Horticulture

The pesticide products used in organic agriculture are generally less toxic, but their reduced efficacy could drive higher application rates, which makes the overall environmental impact of organic agriculture less obvious. In fact, certain pesticides used in organic agriculture have been found to be more toxic than conventional pesticides targeting the same pest . In Läpple and Van Rensburg , the authors found that farmers who entered organic production after the supporting policy was launched are more likely to be profit-driven and less environmentally concerned than farmers who began organic production before any supporting policy was in place. Therefore, studying pesticide use in organic agriculture and how it changes can expand the understanding of organic agriculture and its future. The consolidation into larger operations is another important issue for organic agriculture because it could undermine the perception of organic agriculture as environmentally friendly. Although both the number of organic farms and total organic acreage has increased, blueberry package consolidation still exists if large farms grow faster than small farms.

Meanwhile, the consolidation process had been clearly documented for the organic food processing sector and U.S. agriculture in general . Farm size, measured in acreage, was found to be positively correlated with pesticide use for staple crop productions in the previous literature for conventional agriculture . If this relationship also applies to organic agriculture, then cropland consolidation could have a negative impact on the environment, which means that organic agriculture could become less environmentally friendly than it used to be as the consolidation proceed. I find that the farm size is positively associated with use of sulfur and fixed copper pesticides in the organic crop production. Organic agriculture in California has a diverse crop portfolio, which affects farm size and pesticide use simultaneously. Certain crops are produced in a large scale, measured in acreage, and require intensive pesticide use. How changes in the crop mix interact with the consolidation process is another issue investigated in this essay. The objective of this essay is threefold: to identify organic fields in the PUR database using historical pesticide use records; to characterize the patterns and trends of production and pesticide use for those identified organic fields collectively by crop, crop acreage, year, farm size, and other attributes; to assess the environmental impacts of pesticide use in organic agriculture and the consolidation of organic cropland.

For each agricultural pesticide application, the PUR database specifies the location of application in the variable “COMTRS”, which stands for the county, meridian, township, range, and section as defined by the Public Lands Survey mapping system . This information allows us to locate which section does the field belong and aggregate pesticide usage at the 1×1 mile PLSS section-level, which is the finest spatial scale reported in the PUR database. This detailed section-level analysis of the spatial distribution of organic fields in California and, how it has changed over time, is only possible using my method for identifying organic production fields in the PUR database. In the PUR database, acreage information is recorded as both treated acreage and planted acreage. The former represents the acres physically treated in a pesticide application while the latter remains constant for the field within a year. However, researchers have demonstrated that planted acreage in the PUR database is not consistently reliable for annual crops . So, in this essay, we use the maximum treated acreage in a given year as the acreage for each field for annual crops. This approach assumes that the entire field is treated with pesticide at least once per year. If this assumption is invalid, then the planted acreage will be under counted. As presented below, the validity of this approach is supported by the consistency of state-scale crop acreages that are generated from the PUR database with those from other data sources. One caveat of the PUR database for organic production is that since 2000, pesticide products deemed as having “minimum impacts” are no longer required to be registered with CDPR, which exempts them from the pesticide use reporting requirement. A detailed list of these pesticide ingredients can be found in the California Code of Regulations section 6147 .

Most ingredients exempted from registration are natural or naturally-derived products , which could presumably be used in organic agriculture and have impacts on the surrounding environment. However, these exempted ingredients are not widely applied, based on their minimal amounts of usage in the PUR database prior to 2000 when they were still required to be reported. Therefore, this issue is not likely to invalidate the results, especially because the number of fields where only such ingredients were applied before 2000 is small. For convenience, some chemically-related individual active ingredients were grouped together, such as combining the many different strains of Bacillus thuringiensis, which target different insects and are each treated as a distinct AI in the PUR database, into a single “microbial” group. A detailed list of microbials is available in the appendix. The group of “Copper, fixed” includes the summation of copper, copper oxychloride, copper octanoate, copper oxide, and copper hydroxide; and the two forms of copper sulfate .Organic growers are required to comply with a set of crop management standards, regarding seeds and planting stock practices, soil fertility and crop nutrient management, pest, weed, and disease management, and crop rotation among others . The most relevant requirement for this essay is that there is a 36-month transition period between the last application of any prohibited substance under organic regulations and of- ficially recognized organic production. The field identification method relies on this requirement. First, we constructed a list of allowed and prohibited substances based on various sources . Second, we checked each field in the PUR database, to see which AIs were applied over the previous three years. If there were no applications of any prohibited ingredients, then the field was considered organic as of that year. Organic growers who do not use any chemical tools at all to manage pests and weeds are missing from the PUR database entirely, and therefore not identified in this essay. However, blueberry packaging based on acreage comparisons between the PUR database and other data sources, those growers appear to operate a very limited number of acres. A field could comply with the pest, weed, and disease management standards of the NOP while violating other standards and still not qualify for organic production. Because the PUR database only contains pesticide use information, my method cannot distinguish such fields from actual certified organic fields. On the other hand, growers could follow organic farming practices but choose not to certify their fields for various reasons. However, as mentioned above, the amounts of acreage in these categories must not be very substantial because the PUR-derived organic crop acreages agree with those from CAC compiled sources, suggesting that my method is valid. One caveat of this method is the consistency of field information in the PUR database from year to year. As mentioned previously, the “SITE_LOCATION_ID” on pesticide permits, a number chosen by growers or assigned by county, indicate a physical field location, but the id number may change from year to year. When that ID does change, a new “field” appears in the PUR database for which we do not have information on its historical pesticide applications. In this situation, we assume for annual crops that the land was fallow before a new “SITE_LOCATION_ID” was assigned. This assumption could cause us to overestimate the total organic acreage somewhat, by including fields with a new “SITE_LOCATION_ID” which may have had prohibited substance applications in the past three years. Pasture and rangeland have unique pest management practices and enormous acreage, but they are not covered in this essay as they do not suit my primary purpose of evaluating the environmental impacts of pesticide use in organic crop fields.

For pesticide applications on the identified organic fields in the PUR database, the PURE index was used to assess the potential environmental impacts for five environmental dimensions: surface water, groundwater, soil, air, and pollinator . The PURE index is calculated for five different environmental dimensions: surface water, groundwater, soil, air, and pollinators. For dimensions other than air, index values are calculated based on predicted environmental concentrations and standard toxicity values for relevant organisms. The algorithm used to calculate the predicted environmental concentrations includes the site-specific environmental conditions , which is a major advantage over other indices for assessing the environmental impacts of pesticide use, such as the Environmental Impact Quotient . The predicted environmental concentrations have been proven to align with monitoring data in a previous study . The index value for air is calculated using the predicted volatile organic compound emissions of each pesticide product. Individual index values are normalized to range from 0 to 100 . The PURE index values are calculated for each AI in each pesticide application for each field. These disaggregated index values are then summed at the field level to provide a general index value assessment for each field. To evaluate the overall impact for each crop, an acreage-weighted average across all relevant fields can be determined. This aggregation can be taken one step further, across all AIs, to show the potential impact for all pesticide use using the same aggregation process for organic and conventional fields.Measures of State-specific total organic acreage from different data sources are compared for all seven crops/crop groups available in the USDA organic certification and survey data plus strawberries from the CSC survey data . Among these eight crops/crop groups, four of them are annual and the others are perennial crops. Their organic acreage is plotted in Figure 2.2 and Figure 2.3 below. As mentioned previously, then data sources have different reporting requirements and discrepancies can be caused by a variety of reasons that apply to all crops. In the CDFA registration data, new organic growers report their expected acreage for the next year. If growers decide not to engage in that expected organic production, their registration records remain in the system, which could produce an inaccurate inflation in acreage data, especially for crops that went through a rapid growth of organic production. Growers with less than $5,000 annual organic sales are required to register their production with CDFA but do not have to apply for organic certification. So their acreage might not be counted in USDA data. Both USDA and CDFA data relies on a set of well-defined organic standards and restrictive regulations, which did not exist before 2001. For perennial crops, new registrations with CDFA or new certifications with USDA must include the documentation of orchards before they are fully established . Therefore, we observe new orchards and vineyards in the USDA and CDFA data before they enter the PUR database. If growers adopt the organic pest management program but do not market their products as organic, their acreage is only covered in the PUR database, not the others.Meanwhile, discrepancies in acreage are caused by crop specific reasons. For lettuce , USDA has consistently reported higher acreage values since 2004. In 2016, the lettuce acreage from the USDA organic survey is more than double the acreage from the CDFA registration data . One reason could be potential double or multiple cropping of lettuce in one calendar year. When growers harvested lettuce multiple times from the same field, USDA reported the sum of acres for each harvest while CDFA asked for the size of the field. My method accounts for this phenomenon by counting days between the first and last pesticide applications. Normally, both leaf and head lettuce require at most 130 days from planting to harvest in California . So if two pesticide applications occur more than 130 days apart for the same field, we assume that the lettuce was harvested twice and the acreage would be doubled. After this adjustment, the PUR database still falls short of the acreage documented in the USDA dataset but is in-between the other two sources since 2003. Before 2003, CDFA had more acreage than the other data series because the crop category “lettuce, salad mix”, which contains arugula, red/green mustard, and other crops , used to be reported as lettuce. For strawberries , the CSC data always show somewhat less acreage because their data are derived from surveys and the survey response rate is not reported. For apples , the organic acreage is small compared other perennial crops, which amplifies the potential measurement errors.

The paper concludes with a short summary and plans for the future.The ARIANNA experiment is an array of autonomous radio stations located in Antarctica. Stations have operated at sea-level on the Ross Ice Shelf in Moore’s Bay, about 110 km from McMurdo Station, which is the largest research base on the continent. In addition, two stations have operated at the South Pole, which is colder and higher in elevation than the environment at Moore’s Bay Several architectures were implemented in the prototype array at Moore’s Bay. Most stations consisted of four downward facing log periodic dipole antennas to specifically look for neutrino events, as shown in figure 1. Two other stations at Moore’s Bay and two at the South Pole were configured with eight antennas, which included a mixture of LPDAs and dipoles. These stations were simultaneously sensitive to cosmic rays that interact in the atmosphere and neutrinos. The radio signals are digitized and captured using a custom-made chip design known as the SST. The analog trigger system of ARIANNA imposes requirements on individual wave forms; a high and low threshold must occur within 5 ns, nursery grow bag and multiple antennas channels must meet the high-low threshold within a 30 ns coincidence window.

These criteria are based on the expectation that thermal noise fluctuations are approximately independent, whereas neutrino signals produce correlated high-low fluctuations in a given antenna, and produce comparable signals in multiple antenna channels. These requirements reduce the rate of thermal noise triggers for a given trigger threshold while maintaining the sensitivity to Askaryan pulses from high-energy neutrinos. Once a station has triggered, the digitized wave forms of every antenna channel contain 256 samples with a voltage accuracy of 12 bits. The event size in an eight-channel station is 132 kbits. The waveform data from all channels are piped into an Xilinx Spartan 4 FPGA, and then further processed and stored to an internal 32 GB memory card by an MBED LPC 1768 microcontroller. There are up to eight channels on each board that process the radio signal from each antenna. Once a triggered event is saved to local storage it is then transferred to UC Irvine through a long-range WiFi link during a specified communication window. The ARIANNA stations also use Iridium satellite network as a backup system. Satellite communication is relatively slow, with a typical transfer rate of one event every 2–3 minutes. For both communication methods, currently the hardware system is limited to either communication or data collection. Therefore, neutrino search operations are disabled during data communication.

As radio neutrino technologies move beyond the prototype stage, the relatively expensive and power consumptive AFAR system will be eliminated. Perhaps it will be replaced by a better wireless system, such as LTE, for sites relatively close to scientific research bases, but for more remote locations, only satellite communications such as Iridium are feasible. Given the current limitation of 0.3 events/min imposed by Iridium communication, and the fact that neutrino operations cease during data transfer which generates unwanted deatime, stations that rely solely on Iridium communication are expected to operate at trigger rates from ∼ 0.3 mHz to keep losses due to data transfer, trans, below 3%. The trigger thresholds of ARIANNA are adjusted to a certain multiple of the Signal to Noise Ratio , defined here as the ratio of the maximum absolute value of the amplitude of the waveform to the RMS noise. Currently, the pilot stations are set to trigger above 4.4 SNR to reach the constrained trigger rate of order 1 mHz. In the next section, the expected gain in sensitivity is studied for a lower threshold of 3.6 SNR, which corresponds to 100Hz, the maximum operation rate of the stations. For more information on the ARIANNA detector, see.The real-time rejection of thermal noise that is presented in this article would enable the trigger threshold to be lowered significantly — thus increasing the detection rate of UHE neutrinos — while keeping a low event rate of a few mHz.

To estimate the increase in sensitivity, the effective volume of an ARIANNA station is simulated for the two trigger thresholds corresponding to a thermal noise trigger rate of 10 mHz , and a four orders-of-magnitude higher trigger rate . We use the relationship between trigger threshold and trigger rate from to calculate the thresholds. NuRadioMC is used to simulate the sensitivity of the ARIANNA detector at Moore’s Bay. The expected radio signals are simulated in the ARIANNA detector on the Ross ice shelf, i.e., an ice shelf with a thickness of 576 m and an average attenuation length of approx. 500 m, and where the ice-water interface at the bottom of the ice shelf reflects radio signals back up with high efficiency. The generated neutrino interactions are distributed uniformly in the ice around the detector with random incoming directions. The simulation is performed for discrete neutrino energies and includes a simulation of the full detector response and the trigger algorithm as described above. The resulting gain in sensitivity is shown in figure 2 and increases by almost a factor of two at energies of 1017 eV. The improvement decreases towards higher energies because fewer of the recorded events are close to the trigger threshold but at 1018 eV there is still an increase in sensitivity of 40%.To implement a deep learning filter, the general network structure needs to be optimized for fast and accurate classification. For accuracy, the two metrics are neutrino signal efficiency and noise rejection factor , where ratio is the ratio of correctly identified noise events to the total number of noise events. The goal is to reject several orders-of-magnitude of thermal noise fluctuations while retaining most of the neutrino signals. In the following, the target is 5 orders-of-magnitude thermal noise rejection while providing a high signal efficiency at or above 95%. Typically using a more complex network structure yields more accurate results, but this also creates a slower network. These two constraints need to be optimized as the deep learning architecture is developed. In the following two sections, deep learning techniques are used to train models then study their efficiency and processing time. In section 5.1, a commonly used method of template matching will be investigated to compare with the deep learning approach.NuRadioMC is used to simulate a representative set of the expected neutrino events for the ARIANNA detector, following the same setup as described in section 2.2 but for randomly distributed neutrino energies that follow an energy spectrum expected for an astrophysical and cosmogenic neutrino flux; the astrophysical flux measurement by IceCube with a spectral index of 2.19 is combined with a model for a GZK neutrino flux based on Auger data for a 10% proton fraction. The resulting radio signals are simulated in the four LPDA antennas of the ARIANNA station by convolving the electric-field pulses with the antenna response, plastic growing bag and the rest of the signal chain is approximated with an 80 MHz to 800 MHz band-pass filter. An event is recorded if the signal pulse crossed a high and a low threshold of 3.6 times RMS noise within 5 ns in at least two LPDAs within 30 ns. At such a low trigger threshold, noise fluctuations can fulfil the trigger condition at a non-negligible rate. Therefore, the signal amplitude is required to be at least 2.8 times the RMS noise before adding noise to avoid spurious triggers on thermal-noise fluctuations. In total 121,597 events that trigger the detector are generated and this is called the signal data set in the following. The training data set for thermal noise fluctuations is obtained by simulating thermal noise in the four LPDA antennas and saving only those events where a thermal noise fluctuation fulfills the trigger condition described above. In total 1.1 million events are generated and this is called the noise data set in the following. The limitations of the simulations and their impact on the obtained results are discussed at the end of this article.All of the networks are created with Keras, a high-level interface to the machine-learning library TensorFlow. Our primary motivation is to develop a thermal noise rejection method that operates on the existing ARIANNA hardware with an evaluation rate of at least 50 Hz, which is a factor of 104 larger than our current trigger rate. To increase the execution rate of the neural network, the hardware is one option to optimize; however, any alteration to the hardware is constrained by two main factors: the power consumption of the component and the reliability in the cold climate.

Thus, this study will focus primarily on optimizing the execution rate by identifying the smallest network that reaches our objective. While the number of trainable parameters can give an indication of network size, the number of Floating Point Operations is the chosen metric for network size in this paper. The number of FLOPs can be approximated by multiplying the amount of operations performed by floating point numbers with the amount of nested loop iterations required to classify incoming data. Besides making the network size smaller, another way to improve the network speed is to reduce the input data size. Instead of feeding the signal traces from all four antennas into the network, one way to cut down on the size of input data is to use only the two antennas that caused the trigger. As each signal trace consists of 256 samples, the total input size to the network is 512 samples. In addition, a further reduced input data set is studied for various sizes by selecting the antenna with the highest signal amplitude and only using a window of values around the maximum absolute value. The window size was not fully optimized, but a good balance between input data size and efficiency is 100 samples around the maximum value. The reasoning for this is that the dominant neutrino signal does not span over the whole record length and typically only spans over less than 50 samples. The two network architectures studied in the following are a fully connected neural network and a convolutional neural network , depicted in figure 3. The FCNN used in this baseline test is a fully connected single hidden layer network with a node size of 64 for the 100 input samples and 128 for the 512 input samples, a ReLU activation, and a sigmoid activation in the output layer. The CNN structure consists of 5 filters with 10×1 kernels each, a ReLU activation, a dropout of 0.5, a max pooling with size 10×1, a flattening step to reshape the data, and a sigmoid activation in the output layer. Both the CNN and FCNN are trained using the Adam optimizer with varying learning rates from 0.0005-0.001 depending on which value works best for each individual model. The training data set contains a total of 100,000 signal events and 600,000 noise events, where 80% is for training and 20% is to validate the model during training. Once the network is trained, the test data are used which contain 21,597 signal events and 500,000 noise events. With the sigmoid activation in the output layer, the classification distribution falls between 0 and 1, where close to 0 is noise-like data and close to 1 is signal-like data. Once trained, with the 100 input sample CNN mentioned above, the distribution shown in figure 4 is obtained. From this distribution, the amount of signal efficiency vs. noise rejection can be varied by choosing different network output cut values. Training and testing these networks with each input data size yields the signal efficiency vs. noise rejection plot in figure 5. Each data point corresponds to a different network output value, and the final cut value is chosen by optimizing the noise rejection for the desired signal acceptance. All of these input data sizes produce efficiencies above the required threshold of 95% for signal, and all were able to reach at least 5 orders-of-magnitude noise rejection. Since all of the networks have efficiencies above our target of 95% for signal at 105 noise rejection, the main consideration is the amount of FLOPs required for each network because this directly impacts the processing time. Typically, CNN’s have less parameters overall due to their convolutional nature, which focuses on smaller features within a waveform; comparatively, the FCNN considers the whole waveform to make its prediction, so it requires more node connections.

Epigenetics has been proposed as crucial in shaping plant phenotypic plasticity, putatively explaining the rapid and reversible alterations in gene expression in response to environmental changes. This fine-tuning of gene expression can be achieved through DNA methylation, histone modifications and chromatin remodeling . Small non-coding RNAs are ubiquitous and adjustable repressors of gene expression across a broad group of eukaryotic species and are directly involved in controlling, in a sequence specific manner, multiple epigenetic phenomena such as RNA-directed DNA methylation and chromatin remodeling and might play a role in genotype by environment interactions. In plants, small ncRNAs are typically 20–24 nt long RNA molecules and participate in a wide series of biological processes controlling gene expression via transcriptional and post-transcriptional regulation . Moreover, small RNAs have been recently shown to play an important role in plants environmental plasticity . Fruit maturation, the process that starts with fruit-set and ends with fruit ripening , has been largely investigated in fleshy fruits such as tomato and grapevine. These studies highlighted, among others, plastic planters the vast transcriptomic reprogramming underlying the berry ripening process , the extensive plasticity of berry maturation in the context of a changing environment , and the epigenetic regulatory network which contributes to adjust gene expression to internal and external stimuli .

In particular, small RNAs, and especially microRNAs , are involved, among others, in those biological processes governing fruit ripening . In this work, we assessed the role of small ncRNAs in the plasticity of grapevine berries development, by employing next-generation sequencing. We focused on two cultivars of Vitis vinifera, Cabernet Sauvignon, and Sangiovese, collecting berries at four different developmental stages in three Italian vineyards, diversely located. First, we described the general landscape of small RNAs originated from hotspots present along the genome, examining their accumulation according to cultivars, environments and developmental stages. Subsequently, we analyzed miRNAs, identifying known and novel miRNA candidates and their distribution profiles in the various samples. Based on the in silico prediction of their targets, we suggest the potential involvement of this class of small RNAs in GxE interactions. The results obtained provide insights into the complex molecular machinery that connects the genotype and the environment.RNA extraction was performed as described in Kullan et al. . Briefly, total RNA was extracted from 200 mg of ground berries pericarp tissue using 1 ml of Plant RNA Isolation Reagent following manufacturer’s recommendations. The small RNA fraction was isolated from the total RNA using the mirPremier R microRNA Isolation kit and dissolved in DEPC water. All the steps suggested in the technical bulletin for small RNA isolation of plant tissues were followed except the “Filter Lysate” step, which was omitted.

The quality and quantity of small RNAs were evaluated by a NanoDrop 1000 spectrometer , and their integrity assessed by an Agilent 2100 Bioanalyzer using a small RNA chip according to the manufacturer’s instructions. Small RNA libraries were prepared using the TruSeq Small RNA Sample Preparation Kit , following all manufacturers’ instructions. Forty-eight bar-coded small RNA libraries were constructed starting from 50 ng of small RNAs. The quality of each library was assessed using an Agilent DNA 1000 chip for the Agilent 2100 Bioanalyzer. Libraries were grouped in pools with six libraries each . The pools of libraries were sequenced on an Illumina Hiseq 2000 at IGA Technology Services . The sequencing data were submitted to GEO–NCBI under the accession number GSE85611.In order to investigate whether the overall distribution and accumulation of small RNA is affected by the interaction between different V. vinifera genotypes [Cabernet Sauvignon and Sangiovese ] and environments [Bolgheri , Montalcino and Riccione ], we investigated the regions in the grapevine genome from where a high number of small RNAs were being produced , by applying a proximity-based pipeline to group and quantify clusters of small RNAs as described by Lee et al. . The nuclear grapevine genome was divided in 972,413 adjacent, non-overlapping, fixed-size windows or clusters. To determine the small RNA cluster abundance, we summed the hits-normalized-abundance values of all the small RNAs mapping to each of the 500 bp clusters, for each library .

To reduce the number of false positives, we considered a cluster as expressed when the cluster abundance was greater than the threshold for a given library, eliminating regions where few small RNAs were generated, possibly by chance. Libraries from bunch closure, representing green berries, and 19 ◦Brix representing ripened berries, where used in this analysis. From the 972,413 clusters covering the whole grapevine genome, 4408 were identified as expressed in at least one sample. As showed in Figure 1, CS-derived libraries have a higher number of expressed clusters when compared to SG-derived libraries of the same developmental stage and from the same vineyard. The exceptions were the Sangiovese green berries collected in Riccione and Sangiovese ripened berries collected in Montalcino, which have a higher number of expressed clusters than the respective CS ones. The two cultivars show a completely different small RNA profile across environments. When Cabernet berries were green, a higher number of sRNA-generating regions were found active in Bolgheri than in Montalcino and Riccione. Differently, ripened berries had the highest number of sRNA producing regions expressed in Riccione, while Bolgheri and Montalcino show a similar level of expressed clusters . Sangiovese green berries instead show the highest number of active sRNA-generating regions in Riccione, and this number is twice the number found in Bolgheri and Montalcino that is similar. Ripened berries collected in Montalcino and Riccione show almost the same high level of sRNA-generating clusters, whereas those collected in Bogheri present a lower number . We also noted that when cultivated in Bolgheri, neither Cabernet Sauvignon or Sangiovese change dramatically the number of expressed clusters during ripening, while in Riccione Cabernet Sauvignon shows a 2-fold increase of sRNAproducing clusters, which is not observed in Sangiovese. Next, the small RNA-generating clusters were characterized on the basis of the genomic regions where they map, i.e., genic, intergenic and transposable elements. In general, when the berries were green, the numbers of sRNA-generating loci located in genic and intergenic regions were roughly equal in all environments and for both cultivars, except for Sangiovese berries collected in Riccione, plastic nursery plant pot which show a slight intergenic disposition of sRNA-producing regions . Differently, in ripened berries on average 65% of the sRNA-generating loci were in genic regions, indicating a strong genic disposition of the sRNA-producing clusters . The shift of sRNA-producing clusters from intergenic to mostly genic is more pronounced in Cabernet Sauvignon berries collected in Riccione, with an increase of approximately 20% of expressed clusters in genic regions when berries pass from the green to the ripened stage. When comparing the clusters abundance among libraries, we found that 462 clusters were expressed in all libraries. The remaining 3946 expressed clusters were either shared among groups of libraries or specific to unique libraries. Interestingly, 1335 of the 4408 expressed clusters were specific to Riccione-derived libraries . The other two environments showed a much lower percentage of specific clusters, 263 and 140 in Bolgheri and Montalcino respectively . Comparing the expressed clusters between cultivars or developmental stages, we did not observe a similar discrepancy of specific clusters toward one cultivar or developmental stage; roughly the same proportion of specific clusters was found for each cultivar and for each developmental stage .

Among the 1335 specific clusters of Riccione, 605 were specific to Cabernet Sauvignon ripened berries of and 499 to Sangiovese green berries. Other smaller groups of expressed clusters were identified as specific to one cultivar, one developmental stage or also one cultivar in a specific developmental stage. When comparing the expressed clusters with the presence of transposable elements annotated in the grapevine genome , we noticed that approximately 23% of the sRNA-generating regions were TE-associated. Sangiovese green berries from Riccione have the highest proportion of TE-associated expressed clusters, while Cabernet Sauvignon ripened berries also from Riccione show the lowest proportion of TE associated expressed clusters. Sangiovese berries have the highest percentage of expressed clusters located in TE when cultivated in Riccione, compared to the other two vineyards. Interestingly, Cabernet Sauvignon berries show the lowest proportion of TE-associated clusters when growing in Riccione , independently from the berry stage. In all the libraries, Long Terminal Repeat retrotransposons were the most represented TE. More specifically, the gypsy family was the LTR class associated with the highest number of sRNA hotspots. The other classes of TE associated with the sRNA-generating regions can be visualized in Figure 3B.To determine the global relationship of small RNA-producing loci in the different environments, cultivars and developmental stages, we performed a hierarchical clustering analysis. As showed in Figure 4, the libraries clearly clustered according to the developmental stage and cultivar and not according to the environments. Ripened and green berries had their profile of sRNA-generating loci clearly distinguished from each other. Inside each branch of green and ripened samples, Cabernet Sauvignon and Sangiovese were also well separated, indicating that, the cultivar and the stage of development in which the berries were sampled modulate the outline of sRNA-producing loci more than the environment. Notwithstanding the evidence that developmental stage and variety have the strongest effect in terms of distinguishing samples clustering, we were interested to verify the environmental influence on small RNA loci expression in the two cultivars. Thus, for each sRNA-generating cluster we calculated the ratio between cluster abundance in Cabernet Sauvignon and Sangiovese in each environment and developmental stage, thereby revealing the genomic regions with regulated clusters, considering a 2-fold change threshold, a minimum abundance of 5 HNA in each library and a minimum sum of abundance of 30 HNA . Figure 5 shows how different environments affect the production of small RNAs. In Bolgheri, regardless the developmental stage there were many clusters with a very high abundance level in Cabernet Sauvignon . In Montalcino and even more in Riccione we also observed differences between the expressions of clusters in the two cultivars, with ripened and green berries showing an almost opposite profile in terms of number of clusters more expressed in Cabernet Sauvignon or Sangiovese. When the berries were green, in Montalcino Cabernet Sauvignon shows the highest number of up-regulated clusters, while in Riccione, Sangiovese has the highest number of up-regulated clusters. The opposite behavior was noticed in ripened berries, with Sangiovese having the highest number of up-regulated clusters in Montalcino and Cabernet Sauvignon in Riccione . Notably, we observed a small percentage of regulated clusters exhibiting at least a 10-fold higher abundance of small RNA in Cabernet Sauvignon or Sangiovese when compared to each other . An examination of those clusters showed that a substantial difference could exist between the cultivars, depending on the vineyard and the developmental stage. For example, in Riccione, a cluster matching a locus encoding a BURP domain-containing protein showed a fold change of 390 when comparing green berries of Sangiovese with Cabernet Sauvignon. Similarly, the majority of the highly differentially expressed clusters showed a similar profile: strong bias toward 21-nt sRNAs and a low strand bias. These findings suggest that these small RNAs might be the productof RDR polymerase activity rather than degradation products of mRNAs.We applied a pipeline adapted from Jeong et al. and Zhai et al. to identify annotated vvi-miRNAs, their variants, novel species-specific candidates and, when possible, the complementary 3p or 5p sequences. Starting from 25,437,525 distinct sequences from all the 48 libraries, the first filter of the pipeline removed sequences matching t/r/sn/snoRNAs as well as those that did not meet the threshold of 30 TP4M in at least one library or, conversely, that mapped in more than 20 loci of the grapevine genome . Only sequences 18–26-nt in length were retained. Overall, 27,332 sequences, including 56 known vvi-miRNAs, passed through this first filter and were subsequently analyzed by a modified version of miREAP as described by Jeong et al. . miREAP identified 1819 miRNA precursors producing 1108 unique miRNA candidates, including 47 known vvi-miRNA. Next, the sequences were submitted to the third filter to evaluate the single-strand and abundance bias retrieving only one or two most abundant miRNA sequence for each precursor previously identified. A total of 150 unique miRNA corresponding to 209 precursors were identified as candidate miRNAs.

These are ACC oxidase, which is involved in ethylene biosynthesis; a lipoxygenase, part of a fatty acid degradation pathway giving rise to flavor alcohols such as hexenol; α-expansin 1, a cell wall loosening enzyme involved in fruit softening, and two terpene synthases, which produce important terpenes that contribute to Cabernet Sauvignon flavor and aroma. The high similarity of these transcript profiles indicates that ethylene biosynthesis and signaling may be involved in the production of grape aroma. Supporting this argument, two recent studies have shown that a tomato ERF TF , falling in the same ERF IX subfamily, has a strong effect on ethylene signaling and fruit ripening. The transcript abundance of AtERF6 in Arabidopsis is strongly increased by ethylene, which is triggered by the MKK9/MPK3/MPK6 pathway. The transcript abundance of VviMKK9 in the Cabernet Sauvignon berries was higher in the skin than the pulp, but there were no significant differences for VviMPK3 or VviMPK6 . This is not too surprising since AtMKK9 activates AtMPK3 and AtMPK6 by phosphorylation. In addition, french flower bucket the transcript abundance of AtERF6 in Arabidopsis increases with ROS, SA, cold, pathogens, and water deficit.

There were no visible signs of pathogen infection in these berries. Additional circumstantial evidence for ethylene signaling in the late stages of berry ripening was that the transcript abundance of many VviERF TFs was significantly affected by berry ripening and/or tissue . The transcript abundance of 129 members from the berries was determined to be above background noise levels on the microarray . The expression profiles of the 92 significantly affected AP2/ERF super family members were separated into six distinct clusters by hierarchical clustering and indicated that this super family had a complex response during berry ripening . The 12 members of Cluster 1 responded similarly in both the skin and pulp, gradually decreasing with increasing °Brix with a large decrease in transcript abundance at the 36.7 °Brix level. Cluster 2 with 14 members, including 8 members of the VviERF6 clade, had much higher transcript abundance in the skin with a sharp peak at 23.2 °Brix. Cluster 3 had similar profiles in both the skin and pulp with a peak abundance at 25° Brix. Cluster 4 with 7 members was a near mirror image of cluster 2, with a sharp valley for transcript abundance in the skin between 23 and 25 °Brix.

Cluster 5 had 36 members with a steady increase in transcript abundance in the pulp but no substantial increase in the skin until 36.7 °Brix. Finally, in Cluster 6, there were 13 members with a higher transcript abundance in skins compared to pulp. Their transcript abundance increased with increasing °Brix level, but decreased in the skin. The transcript abundance of important components of the ethylene signaling pathway characterized in Arabidopsis and presumed to be functional in grape were also affected by °Brix level and tissue . Three different ethylene receptors, VviETR1, VviETR2, and VviEIN4 decreased with °Brix level in the skin, however there was very little or no change in the pulp. Likewise, VviCTR1, another negative regulator of ethylene signaling that interacts with the ethylene receptors, decreased between 22.6 and 23.2 °Brix in both the skin and the pulp. The transcript abundance of the positive regulator, VviEIN2, peaked at 25 °Brix in both the skin and the pulp. AtEIN2 is negatively regulated by AtCTR1 and when it is released from repression, turns on AtEIN3 and the ethylene signaling pathway downstream.

The transcript abundance of VviEIN3 increased with °Brix level, peaking at 25 °Brix in the skin, and was much higher than in the pulp. Although more subtle, its profile was very similar to VviERF6L1. Derepression of the negative regulators and the increase in positive regulators indicated that ethylene signaling was stimulated during this late stage of berry ripening.The transcript abundance of many of the genes involved in the isoprenoid biosynthesis pathway peaked between 23 and 25 °Brix level, particularly in the skin; this stimulation of transcript abundance continued in both the carotenoid and terpenoid biosynthesis pathways . DXP synthase is a key regulatory step in isoprenoid biosynthesis and its profile was similar to VviERF6L1; its transcript abundance was correlated with the transcript abundance of several terpene synthases in the terpenoid biosynthesis pathway . About 50% of the putative 69 functional terpene synthases in the Pinot Noir reference genome have been functionally characterized. Another 20 genes may be functional but need further functional validation or checking for sequencing and assembly errors. On the NimbleGen Grape Whole-Genome array there are 110 probe sets representing transcripts of functional, partial and psuedo terpene synthases in Pinot Noir . It is uncertain how many may be functional in Cabernet Sauvignon.

There were 34 probe sets that significantly changed with °Brix or the °Brix and Tissue interaction effect; 20 of these are considered functional genes in Pinot Noir. Terpene synthases are separated into 4 subfamilies in the Pinot Noir reference genome; they use a variety of substrates and produce a variety of terpenes. Many of these enzymes produce more than one terpene. The top 8 transcripts that peaked in the skin at the 23.2 to 25 °Brix stages were also much higher in the skin relative to pulp . Five of the eight probesets match four functionally-classified genes in Pinot Noir ; these terpene synthases clustered very closely with VviTPS54, a functionally annotated – Linalool/- Nerolidol synthase. VviTPS58, a -geranyl linalool synthase, was also in the cluster. The other two probesets match partial terpene synthase sequences in the Pinot Noir reference genome. The transcript abundance of genes involved with carotenoid metabolism also changed at different °Brix levels and with tissue type . CCDs are carotenoid cleavage dioxgenases and are involved in norisoprenoid biosynthesis. The transcript abundance of VviCCD1 changed significantly with °Brix level and was higher in skin than pulp, except at 36.7 °Brix. Likewise, the transcript abundance of VviCCD4a and VviCCD4b changed significantly with °Brix level, but was higher in the pulp than the skin. The transcript abundance of VviCCD4c significantly increased with °Brix level, but there were no significant differences between tissues. VviCCD1 and VviCCD4 produce β- and α-ionone , geranylacetone , and 6-methyl-5-hepten-2-one in grapes. There were no significant effects on the transcript abundance of VviCCD7. The transcript abundance of VviCCD8 significantly increased with°Brix level and was higher in pulp than skin. Phytoene synthase, which was also increased in the skin compared to the pulp , and VviCCD1, have been associated with β-ionone and β-damascenone biosynthesis. Other important grape flavors are derived from the fatty acid metabolism pathway and lead to the production of aromatic alcohols and esters. The transcript abundance of many genes associated with fatty acid biosynthesis and catabolism changed with °Brix level . In particular the transcript abundance of a number of genes were correlated with the transcript abundance of VviERF6L1 including VviACCase, Acetyl-CoA carboxylase; KAS III ; VviOAT, ; VviFAD8; ; VviLOX2 and VviHPL . The transcript abundance of alcohol dehydrogenases was affected by tissue and °Brix level . Some ADHs are associated with the production of hexenol and benzyl alcohol. Methoxypyrazines give herbaceous/bell pepper aromas. They are synthesized early in berry development and gradually diminish to very low levels at maturity. Nevertheless, humans can detect very low concentrations of these aroma compounds. Four enzymes, VviOMT1, VviOMT2, VviOMT3 and VviOMT4 , synthesize methoxypyrazines. The transcript abundance of VviOMT1 was higher in the pulp than the skin . In addition, bucket flower the transcript abundance of VviOMT1 decreased significantly with °Brix level in the pulp. There were no significant differences in the trancript abundance in the skin or pulp for VviOMT2, VviOMT3 or VviOMT4 . There was a high correlation of the transcript abundance of VviOMT1 in the pulp with 2-isobutyl-3-methoxypyrazine concentrations in the berries . The transcript abundance of VviOMT2, VviOMT3, or VviOMT4 in either skin or pulp was not correlated with IBMP concentrations . This is consistent with the suggestion that the pulp is the main contributor of IBMP in the berry. Our data indicated that VviOMT1 in the pulp may contribute to the IBMP concentration in these berries.Orthologs of RIN and SPL tomato transcription factors, which are known to be very important fruit ripening trancription factors, were at much higher transcript levels in the skin and decline with °Brix level .

The transcript abundance of the VviNOR ortholog in grape was higher in the pulp and increased slightly to peak at 25 °Brix. In addition, the transcript abundance of VviRAP2.3, an inhibitor of ripening in tomato , decreased in the skin with a valley at 23.2 °Brix; it belongs to Cluster 4 of the AP2/ERF super family . Of particular interest was VviWRKY53 [UniProt: F6I6B1], which had a very similar transcript profile as VviERF6L1 . AtWRKY53 is a TF that promotes leaf senescence and is induced by hydrogen peroxide. This is the first report we know of implicating WRKY53 in fruit ripening . AtERF4 induces AtWRKY53 and leaf senescence, so the interactions between WRKY and ERF TFs are complex. WRKY TFs bind to the WBOX elements in promoters and VviERF6L1 has a number of WBOX elements in its promoter . In addition, AtMEKK1 regulates AtWRKY53 and the transcript abundance of VviMEKK1 peaked at 23.2 °Brix in the skin as well. Interestingly, the transcript abundance of both VviERF4 and VviERF8, whose orthologs in Arabidopsis promote leaf senescence, were at their highest level of transcript abundance at the lowest °Brix levels examined in this study .This study focused on the very late stages of the mature Cabernet Sauvignon berry when fruit flavors are knownto develop. Cabernet Sauvignon is an important red wine cultivar, originating from the Bordeaux region of France. It is now grown in many countries. They also can contain herbaceous characters such as green bell pepper flavor that are particulary prevalent in under ripe grapes. Grape flavor is complex consisting not only of many different fruit descriptors, but descriptors that are frequently made up of a complex mixture of aromatic compounds. For example, black currant flavor, in part, can be attributed to 1,8-cineole, 3-methyl-1-butanol, ethyl hexanoate, 2- methoxy-3-isopropylpyrazine, linalool, 4-terpineol, and β- damascenone and major components of raspberry flavor can be attributed to α- and β-ionone, α- and β- phellandrene, linalool, β-damascenone, geraniol, nerol and raspberry ketone. Some common volatile compounds found in the aroma profiles of these dark fruits and berries include benzaldehyde, 1-hexanol, 2-heptanol, hexyl acetate, β-ionone, β-damascenone, linalool, and α-pinene. In a study of Cabernet Sauvignon grapes and wines in Australia, Cabernet Sauvignon berry aromas wereassociated with trans-geraniol and 2-pentyl furan and Cabernet Sauvignon flavor was associated with 3-hexenol, 2-heptanol, heptadienol and octanal. In another comprehensive study of 350 volatiles of Cabernet Sauvignon wines from all over Australia, the factors influencing sensory attributes were found to be complex; in part, norisoprenoids and δ − and γ-lactones were associated with sweet and fruity characteristics and red berry and dried fruit aromas were correlated with ethyl and acetate esters. In Cabernet Sauvignon wines from the USA, sensory attributes were complex also and significantly affected by alcohol level of the wine. Linalool and hexyl acetate were postitively associated with berry aroma and IBMP was positively correlated with green bell pepper aroma. In France, β-damascenone was found to contribute to Cabernet Sauvignon wine aroma. Thus, flavor development in berries and wines is very complex, being affected by a large number of factors including genetics, chemistry, time and environment. In this paper we begin to examine the changes in transcript abundance that may contribute to flavor development. We show that the transcript abundance of many genes involved in fatty acid, carotenoid, isoprenoid and terpenoid metabolism was increased in the skin and peaked at the °Brix levels known to have the highest fruit flavors . Many of these are involved in the production of dark fruit flavors such as linalool synthases, carotenoid dioxygenases and lipoxygenases. These genes serve as good candidates for berry development and flavor markers during ripening. A broader range of studies from different cultivars, locations and environments are needed to determine a common set of genes involved in berry and flavor development. A similar study was conducted on the production of volatile aromas in Cabernet Sauvignon berries across many developmental stages, including a detailed analysis of the °Brix levels that was surveyed in this study. They found that the production of alcohol volatiles from the lipoxygenase pathway dominated in the later stages of berry ripening and suggested that the activity of alcohol dehydrogenases also could play an important role.

In certain cases, moir´e heterostructures host super lattice minibands with narrow bandwidth, placing them in a strongly interacting regime where Coulomb repulsion may lead to one or more broken symmetries. In several such systems, the underlying bands have finite Chern numbers, setting the stage for the appearance of anomalous Hall effects when combined with time-reversal symmetry breaking. Notably, in twisted bilayer graphene low current magnetic switching has been observed, though consensus does not exist on the underlying mechanism.The δBI dips may be understood as a consequence of current-driven domain wall motion. As established above, applied current drives nucleation of minority magnetization domains. Once these domains are nucleated, increasing the current magnitude is expected to enlarge them through domain wall motion. Where domain walls are weakly pinned, a small increase in the current δI drives a correspondingly small motion δx of the domain wall, producing a change in the local magnetic field δBI characterized by a sharp negative peak at the domain wall position . We may then use this mechanism to map out the microscopic evolution of domains with current. Fig. 6.5h shows a spatial map of δBI , procona florida container measured at three different values of ISD corresponding to distinct features in the transport data.

Evidently, the domain wall moves from its nucleation site on the device boundary towards the device bulk. Local measurements of δBI as a function of ISD show that this motion is itself characterized by threshold behavior, corresponding to the domain wall rapidly moving between stable pinning sites. A full correspondence of transport features and local domain dynamics is presented in the associated publication. The symmetry of the observed magnetic switching is suggestive of a spin or valley Hall effect driven mechanism. To investigate this hypothesis experimentally, we use local magnetic imaging to directly probe the current-driven accumulation of magnetic moments throughout the density- and displacement field tuned phase space. Figs. 6.6c-e show δBI maps measured at three different points, away from the regime where the ground state is ferromagnetic. These fillings correspond to the Hubbard band edges, where the Berry curvature is expected to be enhanced by the appearance of correlation driven gaps. Notably, large spin Hall effects are observed near ν = −1 even far from band inversion, with possible implications for the nature of the strong insulating state observed there. We have shown here that the combination of intrinsic spin Hall effect with intrinsic magnetism provides a mechanism for a current-actuated magnetic switch in a single two dimensional electron system. The physical properties we invoke to explain this phenomenon are generic to all intrinsicChern magnets.

We emphasize that in both twisted bilayer graphene and our current MoTe2/WSe2 heterostructure, magnetic switching arises in regimes for which doping, elevated temperature, or disorder ensure that electrical current flows in the sample bulk. Ultra-low current switching of magnetic order has also been observed in twisted bilayer graphene. In that system, where spin orbit coupling is negligible, nearly identical mechanisms may arise mediated by orbital, rather than spin, Hall effects. The bulk nature of the spin Hall torque mechanism means that similar phenomena should manifest not only in the growing class of intrinsic Chern magnets, but in all metals combining strong Berry curvature and broken time-reversal symmetry, including crystalline graphite multi-layers. Research into charge-to-spin current transduction has identified a set of specific issues restricting the efficiency of spin torque switching of magnetic order. Spin current is not necessarily conserved, and as a result a wide variety of spin current sinks exist within typical spin torque devices. Extensive evidence indicates that in many spin torque systems a significant fraction of the spin current is destroyed or reflected at the spin-orbit material/magnet boundary. In addition, the transition metals used as magnetic bits in traditional spin-orbit torque devices are electrically quite conductive, and can thus shunt current around the spin-orbit material, preventing it from generating spin current. These issues are entirely circumvented here through the use of a material that combines a spin Hall effect with magnetism, and as a result of these effects this spin Hall torque device has better current-switching efficiency than any known spin torque device.

We started this discussion with a favorable comparison of the impact of disorder on the ABMoTe2/WSe2 Chern magnet to graphene-based Chern magnets. I’m sure the reader was just as disappointed as we were to see the dramatic disorder landscape on display in Fig. 6.4E, which presents a map of the magnetization in the AB-MoTe2/WSe2 Chern magnet. This is not a refutation of our original claims; it remains true that the repeatability of the fabrication protocol of the AB-MoTe2/WSe2 Chern magnet is unambiguously much better than that of tBLG/hBN, or even tMBG. It is also easy to lose track of the scale of these images- the tBLG/hBN Chern magnet was only a few square microns, whereas this sample supports a Chern magnet that is almost a hundred square microns in area. The presence of these ‘holes’ in the magnetization of this Chern magnet is not a result of strong twist angle disorder.We do not know the precise origin of these holes, but there are a few possibilities that we can discuss. Bubbles are some of the most common defects in stacks of two dimensional crystals, and they can form between any two layers of a stack. As presented and described in Fig. 6.7A-C, AFM imaging reveals topographic defects precisely aligned with the regions in the Chern magnet in which magnetism has been destroyed. There are two clearly distinct distributions of defects, with thicknesses that differ by about an order of magnitude. It is possible that these correspond to bubbles between two distinct pairs of layers of the stack. Another possibility is that partial oxidation or deliquescence of the MoTe2 crystal has occured. This crystal is indeed air and moisture sensitive, and degradation can happen even inside a glovebox, as illustrated for a CrI3 crystal in Fig. 6.7D-F. Whatever issue is generating this disorder, it will likely be necessary to resolve it in order to fabricate more sophisticated devices based on this Chern magnet.We have so far discussed a variety of phenomena realized in gate-tunable exfoliated heterostructures. In all cases, these phenomena were accessible experimentally because of the presence of a moir´e super lattice, which gave us access to electronic bands that could be completely filled or depleted at will using an electrostatic gate. We will next be discussing an atomic crystal without a moir´e super lattice. This material does not have flat bands, and we will have no hope of completely filling or depleting any of the bands in the system. Instead, it has features in its band structure that lend themselves to interaction-driven phenomena, procona London container specifically flat-bottomed bands satisfying the Stoner criterion. The material we will be studying is an allotrope of three-layer graphene called ABC trilayer graphene. In addition to a variety of other interesting phases, this material supports both spin and orbital magnetism. We will discuss why this is the case, and we will study the ABC trilayer magnets using the nanoSQUID microscope.As in three dimensional crystals, many two dimensional crystals have multiple allotropes that are stable under different conditions. Trilayer graphene is such a material. We label multilayer grapheneallotropes using letters that refer to the relative positions of atoms of different layers, projected onto the two dimensional plane. We have already encountered ABA trilayer graphene in the introduction, and this material has atoms in the third layer aligned to atoms in the first layer. At room temperature and pressure the ABA stacking order is preferred, but trilayer graphene has a metastable allotrope, ABC trilayer graphene, that can either be prepared or found naturally occurring. In ABC trilayer graphene atoms in the third layer are aligned neither with the first nor with the second layer.

ABC trilayer graphene has band structure that differs significantly from ABA trilayer graphene, and these differences have important consequences for its properties.The band structure of ABC trilayer graphene at two different displacement fields is illustrated in Fig. 7.1. In the absence of a displacement field, the system is metallic at all electron densities. When a large displacement field is applied to the system, it becomes a band insulator when the Fermi level is tuned between the two resulting bands. This is the regime of displacement field that we will be discussing. ABC trilayer graphene has extremely weak spin-orbit coupling, so the spin degree of freedom is present and more or less completely orthogonal to electronic degrees of freedom, contributing only a twofold degeneracy to the band structure. Just like most other allotropes of graphene, ABC trilayer graphene has valley degeneracy, and this produces an overall fourfold energetic degeneracy of its band structure. This is illustrated in Fig. 7.2. As is abundantly clear from these plots, the bandspresent in ABC trilayer graphene are not flat; they have extremely large bandwidths. However, the bands do satisfy the flat-bottomed band condition, and as a result we can expect these systems to be able to spin- and valley-polarize without paying significant kinetic energy costs.A schematic of the device we will discuss is presented in Fig. 7.3A. This device allows us to perform several different experiments: we can tune the electron density and displacement field in the ABC trilayer graphene layer, we can measure in-plane electronic transport , and we can measure the out-of-plane capacitive conductivity as a function of electron density and displacement field. Data extracted from this contrast mechanism is presented in Fig. 7.3B. This dataset is restricted to the hole band; i.e., the bottom band in all of the plots we have so far encountered. Sharp features in this dataset correspond to spontaneous symmetry breaking; these features are marked with the numbers and . The right side of this plot, labelled with an electron density of zero, corresponds to charge neutrality in this system and lies in the gap of the band insulator. Therefore and both correspond to situations in which the hole band is very slightly filled. The valley and spin subbands of ABC trilayer graphene are presented in schematic form in Fig. 7.4A in the absence of electronic interactions. When we tune the Fermi level into these bands and activate interactions, we cannot produce a gap- the bandwidths of these bands are far too high- but we can produce full spin or valley polarization, as illustrated in Fig. 7.4B. The precise situations in which we find this system at and are presented in Fig. 7.4C and D; these situations correspond repsectively to full spin polarization but no valley polarization in and full spin and valley polarization in . Valley polarization couples strongly to transport, generating a large anomalous Hall effect and ferromagnetic hysteresis, as presented in Fig. 7.4E.Although these magnets occur in an atomic crystal, they are composed entirely of electrons we have forced into the system with an electrostatic gate, and as a result we can expect their magnetizations to be considerably smaller than fully spin-polarized atomic crystals. We will use the nanoSQUID microscope to image these magnetic phases. An optical image of the ABC trilayer graphene device used to produce data for the publications is presented in Fig. 7.5A. A nanoSQUID image of this region using AC bottom gate contrast is presented in Fig. 7.5B. This magnetic image was taken in the same phase in which we observe magnetic hysteresis, as presented in Fig. 7.4E. Clearly the system is quite magnetized; we also see evidence of internal disorder, likely corresponding to bubbles between layers of the heterostructure. We can park the SQUID over a corner of the device and extract a density- and displacement field-tuned phase diagram of the magnetic field generated by the magnetization of the device; this is presented in Fig. 7.5C. Electronic transport data of the same region is presented in Fig. 7.5D. The spin magnet has only a weak impact on electronic transport, but the valley ferromagnet couples extremely strongly to electrical resistance. The system also supports a pair of superconductors, including a spin-polarized one; these phases are subjects of continued study. Capacitance data over the same region of phase space is presented in Fig. 7.5E.ABC trilayer graphene is the first atomic crystal known to support purely orbital magnetism. Other related systems have since been discovered to host similar phenomena, including bilayer graphene. 

The properties of crystals differ from the properties of atoms floating in free space because the atomic orbitals of the atoms in a crystal are close enough to those of adjacent atoms for electrons to hopbetween atoms. The resulting hybridization of atomic orbitals produces quantum states delocalized over the entire crystal with the capacity to carry momentum. This situation is shown in schematic form in Fig. 1.9A. For quantum states delocalized over the entire crystal, position ceases to be a useful basis. Instead, under these conditions we label electronic wave functions by their momenta, kx and ky. The atomic orbitals that prior to hybridization had discrete energy spectra now have energy spectra given by discrete functions of momentum, f. We call these functions electronic bands.Electrons loaded into the electronic bands of a two dimensional crystal will occupy the quantum states with the lowest available energies, plastic planter pot so we can specify a maximum energy at which we expect to find electrons for any given electron density. We call that energy EF , the Fermi level. We can raise the Fermi level by adding additional electrons to the crystal, as shown in Fig. 1.9B. We have already discussed how two dimensional crystals naturally allow for manipulation of the electron density, and thus the Fermi level.

We have also already discussed how the application of an out-of-plane electric field to a two dimensional crystal will change the structure of the atomic orbitals supported by that crystal. It naturally follows that atomic orbitals so modified will produce different electronic bands, as shown in Fig. 1.9C. It is relatively straightforward to compute how electronic bands will respond to the application of a displacement field . We will be using the momentum and energy basis for the rest of this document; this basis is known as momentum space. The simplest experiment we can perform to probe the electronic properties of a two dimensional crystal in this geometry is an electronic transport experiment, in which a voltage is applied to a region of the crystal with another region grounded, so that electrical current flows through the crystal. We can check if the crystal supports any electrical transport at all, and if it does we can measure the electrical resistance of the crystal this way, in close analogy to how this is done for three dimensional crystals. Crystals will only accept and thus conduct electrons if there are available quantum states at the Fermi level; we call these crystals metals , and they can be identified in band structure diagrams by the intersection of the Fermi level with an electronic band . Crystals without empty quantum states at the Fermi level will not accept and conduct electrons , and they can be identified in band structure diagrams with crystals for which the Fermi level does not intersect with an electronic band .

There exists a variety of other experiments we can perform on two dimensional crystals in order to understand their properties. Two dimensional crystals can support electronic transport in the in-plane direction if they are metals, as shown in Fig. 1.10. Capacitors can also support electronic transport in the out-of-plane direction, as long as that electronic transport occurs at finite frequency. The same structure that we use to modify the electron density and ambient out-of-plane electric field of a two dimensional crystal can also be used as a capacitive AC conductor, as illustrated in Fig. 1.11A. The conductance will depend only on the frequency at which an AC voltage is applied and the geometry of the parallel plate capacitor. However, if a two dimensional crystal is added in series, the capacitance of the top gate to the bottom gate may be substantially modified. If the two dimensional crystal is an insulator, electric fields will penetrate it and the capacitance between the two gates will not change. However, if the two dimensional crystal is a metal it will accept electrons and cancel the applied electric field, dramatically reducing the capacitance between the top and bottom gates and neutralizing the AC current through the capacitor. This technique can be used to measure the electronic properties of specifically the bulk of a two dimensional crystal; it is the property that was both calculated and measured in Fig. 1.2C and D.

These two techniques are the bread and butter of the experimental study of two dimensional crystals, because they require only the ability to create stacks of two dimensional crystals and access to tools common to the study of all other microelectronic systems. However, the primary focus of this thesis will be on systems for which the nanoSQUID microscope can provide important information that is inaccessible to these techniques, and so we will discuss a few such systems next.Consider the following procedure: we obtain a pair of identical two dimensional atomic crystals. We slightly rotate one relative to the other, and then place the rotated crystal on top of the other . The resulting pattern brings the top layer atoms in alignment with the bottom layer atoms periodically, but with a lattice constant that is different from and in practice often much larger than the lattice constant of the original two atomic lattices. We call the resulting lattice a ‘moir´e super lattice.’ The idea to do this with two dimensional materials is relatively new, but the notion of a moir´e pattern is much older, and it applies to many situations outside of condensed matter physics. Pairs of incommensurate lattices will always produce moir´e patterns, and there are many situations in daily life in which we are exposed to pairs of incommensurate lattices, like when we look out a window through two slightly misaligned screens, or try to take pictures of televisions or computer screens with our camera phones. Of course these ‘crystals’ differ pretty significantly from the vast majority of crystals with which we have practical experience, so we’ll have to tread carefully while working to understand their properties. To start with, if we attempt to proceed as we normally would- by assigning atomicorbitals to all of the atoms in the unit cell, computing overlap integrals, and then diagonalizing the resulting matrix to extract the hybridized eigenstates of the system- we would immediately run into problems, because the unit cell has far too many atoms for this calculation to be feasible. Some moir´e super lattices that have been studied in experiment have thousands of atoms per unit cell. There exist clever approximations that allow us to sidestep this issue, and these have been developed into very powerful tools over the past few years, 30 litre plant pots but they are mostly beyond the scope of this document. I’d like to instead focus on conclusions we can draw about these systems using much simpler arguments. The physical arguments justifying the existence of electronic bands apply wherever and whenever an electron is exposed to an electric potential that is periodic, and thus has a set of discrete translation symmetries. For this reason, even though the moir´e super lattice is not an atomic crystal, we can always expect it to support electronic band structure for the same reason that we canal ways expect atomic crystals to support band structure. Two crystals with identical crystal symmetries will always produce moir´e super lattices with the same crystal symmetry, so we don’t need to worry about putting two triangular lattices together and ending up with something else.Another property we can immediately notice is that the electron density required to fill a moir´e super lattice band is not very large.

This can be made clear by simply comparing the original atomic lattice to a moir´e super lattice in real space . Full depletion of a band in an atomic crystal requires removing an electron for every unit cell , and full filling of the band occurs when we have added an electron for every unit cell. We have already discussed how this is not possible for the vast majority of materials using only electrostatic gating, because the resulting charge densities are immense. Full depletion of the moir´e band, on the other hand, requires removing one electron per moir´e unit cell, and the moir´e unit cell contains many atoms . So the difference in charge density between full filling and full depletion of an electronic band in a moir´e super lattice is actually not so great , and indeed this is easily achievable with available technology. Before we go on, I want to make a few of the limitations of this argument clear. There are two things this argument does not necessarily imply: the moir´e bands we produce might not be near the Fermi level of the system at charge neutrality, and the bandwidth of the moir´e super lattice need not be small. In the first case, we won’t be apply to modify the electron density enough to reach the moir´e band, and in the latter, we won’t be able to fill the moir´e band’s highest energy levels using our electrostatic gate. We know of examples of real systems with moir´e super lattice bands that fail each of those criteria. But if these moir´e super lattice bands are near charge neutrality, and if their bandwidths are small, then we should be able to easily fill and deplete them with an electrostic gate.This makes them desirable targets for the types of experiments we’ve discussed above. Finally, moir´e super lattice bands inherit any electronic degeneracies- like, for example, electron spin- that came with the original lattice. We haven’t discussed electronic degeneracies yet, and we will shortly. So if a moir´e super lattice satisfies all of these criteria, then it will provide a set of electronic bands that can be completely filled or depleted with an electronic gate. I’m sure this seems to the reader like a pretty niche system, and that’s more or less because it is. There aren’t too many material systems that need their atomic bonds aligned with a mechanical goniometer, and it’s hard to imagine ever integrating such a procedure into an industrial fabrication line. However, it’s tough to adequately express how hard it would be to replicate the properties of a moir´e super lattice band in an atomic crystal. I made an attempt to do so in the introduction to this thesis; suffice to say the control we have over the properties of these systems is more or less unprecedented within experimental condensed matter physics, and this means that we can perform experiments on electronic phases in these systems that would be difficult or impossible in atomic crystals.A variety of scanning probe microscopy techniques have been developed for examining condensed matter systems. It’s easy to justify why magnetic imaging might be interesting in gate-tuned two dimensional crystals, but magnetic properties of materials form only a small subset of the properties in which we are interested. Scanning tunneling microscopy is capable of probing the atomic-scale topography of a crystal as well as its local density of states, and a variety of scanning probe electrometry techniques exist as well, mostly based on single electron transistors. It’s worth pointing out that if you’re interested specifically in performing a scanning probe microscopy experiment on a dual-gated device, then these techniques both struggle, because the top gate both blocks tunnel current and screens out the electric fields to which a single electron transistor would be sensitive. Magnetic fields have an important advantage over electric fields: most materials have very low magnetic susceptibility, and thus magnetic fields pass unmodified through the vast majority of materials . This means that magnetic imaging is more than just one of many interesting things one can do with a dual-gated device; in these systems, magnetic imaging is a member of a very short list of usable scanning probe microscopy techniques. The simplest way in which we can use our nanoSQUID magnetometry microscope is as a DC magnetometer, probing the static magnetic field at a particular position in space . There are situations in which this is a valuable tool, and we will look at some DC magnetometry data shortly, but in practice our nanoSQUID sensors often suffer from 1/f noise, spoiling our sensitivity for signals at low or zero frequency. One of the primary advantages of the technique is its sensitivity, and to make the best of the sensor’s sensitivity we must measure magnetic fields at finite frequencies. We have already discussed how we can use electrostatic gates to change the electron density and band structure of two dimensional crystals.

Across two training phases participants learned about categories A, B, C, and D in Table 1 via inference or classification. Eye tracking was used throughout to monitor participants’ attention to the three feature dimensions and the category label. A test phase examined classification performance and attention profiles as people made novel category contrasts. From prior research we expected classification subjects to learn to ignore the irrelevant dimensions during training; this attention optimization should lead to a difficulty in making novel classifications. In contrast, prior research has demonstrated a tendency for inference learners to acquire within-category information, suggesting a general motivation to learn about all the dimensions in the inference task. Such motivation can potentially produce flexible category representations—that is—ones that support novel contrasts. Measuring eye movements during training will help explain differences in concept flexibility between groups. In contrast to previous studies comparing inference and classification, black plastic nursery pots a change was introduced to our inference training procedure: One of the dimensions, the contrast dimension 3, was never queried .

This change was made to better equate the two tasks; allowing inference participants to ignore task-irrelevant dimensions just like classification learners could. This allowed a test of whether inference learners are in fact generally motivated to learn about category features, or whether the demands of the task, i.e., querying the features, is what draws learners’ attention.Learning AB and CD training performance are shown in Figure 2. The figure shows average classification performance for the classification group and relevant cue inference performance. Both classification and inference groups improved over training blocks, but classification training was easier than inference training, with a higher proportion correct over blocks. The inference learners performed above chance levels in predicting the valid cue, t = 4.46, p < .01, but were marginally lower than the classification group, t = 1.81, p < .10 on the last AB training block. The CD training blocks were similar. Fixations A crucial question was whether inference learners fixated the non-queried dimension during learning. If inference is a more natural learning task than classification, it should motivate a general interest in learning about the category dimensions; fixations should be distributed to all dimensions, regardless of whether those dimensions are queried. However, if it is the attentional demands of the inference task that drive learning about dimensions , then fixations should shift away from the non-queried dimension, since it is no longer immediately relevant for the task.

The latter result would suggest that differences in what is learned via inference and classification are from different attentional requirements, and not motivational factors. Eye fixations will be used to distinguish between these two possibilities. Figure 3 shows proportion of fixations to category label and dimensions over AB and CD blocks, as a function of task. Replicating our earlier work, at the beginning of learning, the average classification learner fixated dimensions about equally. We also observed the expected shift in fixations from irrelevant to relevant dimensions, until irrelevant dimensions were fixated rarely or not at all. At the onset of CD training in block six, there is uneven attention distribution resulting from the learned fixation patterns from AB training, so that in the first trial of CD training, classification learners were not fixating the contrast dimension or the CD relevant dimension. A second attention optimization obtained for classification subjects. Recall that the contrast dimension was never queried. If inference motivates a general interest in the category features, we should observe continued fixations to the contrast dimension, in spite of it now being task-irrelevant. However, Figure 3 shows that throughout learning, inference learners largely ignored the contrast dimension. Although attention to dimensions 1 and 2 remained high throughout learning, even in the first learning block inference learners largely ignored the contrast dimension.

In fact, in the first block of learning, the amount of time fixating the contrast dimension was already significantly less than that of fixating the other two dimensions and the category label . Apparently, inference learners do in fact optimize their attention away from task-irrelevant cues. Attention optimization in the inference task contradicts the idea that inference motivates a general interest in the category features beyond what is strictly necessary. Rather, the results of Figure 3 support the idea that what distinguishes classification from inference is the attentional demand it places on the learner. Learners fixate dimensions because the task requires it and not because of motivational factors. Any motivation there may have been to learn about all of the category features extinguished quickly . Switch-trial performance Eye fixation data have ruled out that inference motivates general interest in category features. By not querying the contrast dimension in the inference condition, we allowed inference learners the opportunity to optimize their attention, just as the classification learners could. In fact, inference learners optimized their attention to just those queried dimensions, ignoring the never-queried contrast dimension. As a result of this manipulation, the inference learners may now struggle to include the contrast dimension, since they largely ignored it during training. On the other hand, although the inference learners never directed their attention to the contrast dimension, because it was not part of the task, they never had to learn to direct their attention away from that dimension either. Rather, the task focused their attention more on the two queried cues, and inference subject learned which dimensions were task-relevant. It is this fact that may still allow inference learning to nevertheless produce flexible attention allocation. By not learning to ignore the contrast dimension, inference learners may be free to use it during the switch trials. Blocks 10 and 11 of Figure 2 show proportion correct for switch-classification. In spite of not deploying significant fixations to the contrast dimension during training, the inference condition nevertheless showed an advantage during the switch trials. In the first block of switch trials, the inference group outperformed the classification group , t = 1.95, p = .064. Likewise, during the second block of switch trials, the inference group outperformed the classification group , t = 2.26, p < .05. Spending a large amount of time fixating a dimension during learning does not seem necessary for using that dimension later in a flexible way. Whatever inference subjects learned during training allowed them to perform well during switch trials. The eye movement results from training showed that classification and inference learners both largely ignored the contrast dimension. It makes sense then that classification learners should fail to use the contrast dimension during the switch trials, 30 plant pot but what allowed inference learners to have more flexible category representations than the classification group? Figure 4 shows learners’ attention allocation to the contrast dimension as a function of trial for the first block of AB and CD training . The figure shows that at the trial level, the largest attentional difference between the two conditions was that the classification learners allocated more attention to the contrast dimension early in learning. The different patterns of attention reflect different reasons the two groups probably ignored the contrast dimension. Inference learners ignored it because the task directed their attention to those dimensions being queried. Since the contrast dimension was never queried, their attention was never allocated to it. Classification learners were in a different position. From their perspective, any or all dimensions could have been important for getting the answer right, so they had to learn to ignore the contrast dimension, as they gradually discovered that the contrast dimension didn’t help them classify As from Bs or Cs from Ds.

We suspect that this is why there is an initial increase in fixations to the contrast dimension in the first CD block, because classification learners attended to it, and then learned that it was useless in classifying Cs and Ds. Classification learners’ fixation results reflected a learned inattention to the contrast dimension, which probably caused their difficulty in attending to the contrast dimension during the switch trials.We began with the observation that real-life categorizers can make novel category contrasts and that information learned about one set of categories transfers to another without difficulty. This observation seemed to be at odds with the robust finding that people in classification experiments tend to optimize their attention to the fewest necessary dimensions. Such optimization would necessarily force learners to reallocate attention when previously irrelevant dimensions at once become relevant. To resolve the contradiction that people can make novel category contrasts on one hand and but also tend to optimize attention on the other, we looked to other types of learning tasks they may produce classification performance that is less optimal but more flexible overall. Inference training seemed like the best candidate. There were two reasons for this. The first was based on evidence that inference yields a special type of processing in humans; although the exact source of this special processing was until now not entirely clear, classification learning has been found to cause humans to attend to diagnostic information and inference learning can cause learners to focus on within category correlations and prototypical features. We imagined that such differences may reflect that inference is a more typical learning task than classifying, and it isn’t hard to imagine how familiarity in the learning task can lead to greater ease and flexibility in using the acquired information. Our second hypothesis for how inference learning could yield flexible category representations was based on differences in attentional demands of inference and classification. Whereas most classification tasks allow learners to ignore some of the irrelevant dimensions, in the typical inference learning experiment, all of the dimensions are queried several times throughout training. Focusing people’s attention on all of the dimensions in this way may cause people to look at all dimensions on every trial, in order to prepare for future queries. In fact, the eye tracking results from this study show that never querying one of the dimensions allows the inference learner to optimize their attention to only those task-relevant dimensions, i.e., those dimensions that are sometimes queried. As it turned out, our initial hypotheses about inference learning were not exactly right. Our data showed that inference subjects very quickly ignored the never-queried dimension. Significant differences in fixations to the contrast dimension were found within the first learning block. Apparently, attending to the contrast dimension during training was not necessary for creating flexible category representations. Rather, what gave subjects the advantage in switch-classification trials is that they never had to learn to ignore the contrast dimension, as the classification subjects did, as evidence by the much larger drop in attention to the contrast dimension from the beginning to the end of training in the classification condition. In other words, classification subjects were harmed in their task by their learned attention profiles, but the inference subjects were not. Such a finding is in fact consistent with theories of attention and category learning. Several models, for example, RASHNL Kruschke and Johansen , and EXIT Kruschke , which are based on Macintosh’s theory of learned attention, propose that attention weights are learned for a given set of inputs. In these models, if feature inputs are irrelevant, or if for other reasons the features increase the number of classification errors committed, the attention system will direct attention away from those features in favor of others. These attention mechanisms help the models explain a large array of blocking and highlighting phenomena in addition to benchmark category learning data. They also explain why itis that our classification subjects failed to redirect attention to the contrast dimension during the switch trials. Beyond supporting certain theories of categorization and attention, our results underscore an important difference between the attention profiles acquired through trial and error learning and those that arise out of task goals. It seems that ignoring features as a result of discovering that they are statistically irrelevant over numerous trials is qualitatively different than cues that are never queried, and are thus irrelevant for the task. Thus, how the learner acquires an attention profile is as important as the attention profile itself.The 2020–2025 Dietary Guidelines for Americans encourages the intake of a variety of plant-based foods including nuts and berries. With the goal of increasing current knowledge on nuts and berries, as well as addressing research challenges and opportunities, the Nuts and Berries Conference: Pathways to Oxidant Defense, Vascular Function, and Gut Microbiome Changes was held on 5 to 6 May, 2022 at the University of California, Davis. Tree nuts and berries were selected as the focus of the conference for their unique composition, bioactivity, and multitude of associated health-promoting qualities.

The soil was formed in place from the weathering of softly to moderately consolidated granitic sediments. The particle size distribution of the surface soil is 63% sand, 25% silt and 12% clay. At the onset of the experiment, boron tissue levels were in the adequate range, 40 ppm. In both experiments, drip irrigations during the season were based on a schedule using historical evapotranspiration and developed for raisin vineyards in the San Joaquin Valley . The irrigation source was high-quality pump water with a boron concentration less than 0.1 ppm. The experimental design and methods were identical in both vineyards, except that the 1/16-poundper-acre boron treatment was omitted in the second vineyard. The Fresno County trial was discontinued after tissue samples were taken at bloom and veraison in 1998.At both the Tulare and Fresno county sites, boron uptake was rapid when fertilizer was applied in the spring. In both vineyards, applying boron at 2/3 or 1 pound per acre increased the boron concentration in blades by bloom, 3 weeks after application. Boron increased further in blades by veraison . In the Tulare County vineyard, boron in bloom tissue increased from a questionable deficiency range to adequate; at the Fresno location, raspberry container size boron in bloom tissue increased from 40 ppm to 54 ppm, a dramatic increase considering boron fertilizer was applied just 3 weeks prior. This indicates that boron uptake is rapid.

None of the fertigation treatments resulted in either symptoms of boron toxicity or deficiency. Applying boron at 1/3 pound per acre or less did not significantly increase boron tissue levels by bloom or veraison at either site the first year. Fertigation over consecutive years was evaluated at the Tulare County location. Boron in grapevine tissue continued to increase with consecutive years of application. At the higher fertilizer rate , boron levels in blades increased from 35 ppm in control vines to about 60 ppm. We speculate that continuing with annual applications of 1 pound boron per acre would result in toxicity within 4 to 5 years. The 1/3-pound-per-acre rate significantly elevated boron in blades by veraison of the second year to adequate levels . There were no visual signs of toxicity in any of the fertilizer treatments, even when boron was applied at 2 pounds per acre in a single application. Boron levels in tissue remained unchanged 2 years after fertilization was discontinued at the Tulare County location . Rainfall during this experiment was below normal, which helped minimize leaching. Also, well-managed drip irrigation minimizes leaching. Under drip irrigation, salts tend to accumulate near the soil surface and 2 to 3 feet away from the drip line, with minimal water and salt movement below the root zone when irrigations are accurately scheduled . Boron concentrated more in the blades than in the petioles in response to fertilization.

At the onset of the Tulare County experiment, boron concentrations in petioles and blades were similar at 31 ppm and 34 ppm, respectively. Fertilizing with 1 pound boron per acre for 2 consecutive years resulted in a 25% increase of boron in petioles but a 76% increase in blades . Allfertilizer treatments increased boron in blades more than in petioles, indicating that blades should be sampled when monitoring the vines’ boron status following fertilization. Potential boron toxicity values at the time of sampling during the bloom period are 80 ppm for petioles and 120 ppm for blades, and in mid- to late summer are 100 ppm for petioles and 300 ppm for blades.Annual boron fertigation at 1/3 pound per acre elevated grapevine tissue levels from questionable to the adequate range within 2 years . In addition, tissue boron levels remained unchanged 2 years after fertilization was discontinued. This is probably because leaching was reduced by two factors: below-normal rainfall and accurately scheduled drip irrigations. After fertilization, boron was concentrated more in blades than in petioles, indicating that blades are the best choice for monitoring toxicity. Blade samples should be monitored on a routine basis and fertilizer amounts should be adjusted accordingly to avoid boron toxicity or deficiency. The results of this research can be applied to other drip-irrigated vineyards in the San Joaquin Valley under similar conditions: rapidly drained soils, high quality irrigation water, and low boron content in soil, water and vine tissue. In other regions of the state where winter rainfall is much higher, there would presumably be more leaching of boron fertilizer during winter months and less carryover time after fertilization is discontinued. In contrast, less leaching and greater carryover of boron would be expected in areas of less rainfall or on soils with finer texture and higher water-holding capacity.

The amount of boron fertigation used in a maintenance program will vary with leaching potential. These variables underscore the importance of monitoring boron in tissue when developing a long-term fertilization program.Stroke recovery is an exhausting, isolating, and expensive process. Physical therapy to recover limb function and neural pathways is the most expensive part of this process due to the need for frequent one-on-one appointments with physical therapists over the course of months or even years that may not be covered by insurance . In addition, many stroke patients prefer athome rehabilitation whenever possible to allow for more schedule flexibility and to avoid the need to find transportation to what are often extremely distant specialist centers . The need for frequent one-on-one appointments can be reduced through robot-assisted therapy , which improves the quality of both group therapy and at-home physical therapy .Many previous studies have utilized common stroke rehabilitation techniques to create robots to retrain motor movement in the upper limbs by utilizing TST and repetitive motions similar to what a physical therapist might assign as an exercise for a patient . TST can lead to muscle pain and fatigue, however RAT and non-RAT TST have been shown to result in the same level of pain and fatigue for patients . In addition, RAT can include systems to monitor fatigue and pain. Patients also rate higher levels of enjoyment and interest in physical therapy when utilizing these robots, so the only cons in these robots come from their common design traits . For stabilization and ease of calibration, these systems are often large and heavy so the position of the exoskeleton or robotic interface is easily known . This leads to increased costs for consumers in both purchasing the additional bulky casing around the robot as well as creating a space for such personalized gym equipment. In addition, these heavy systems are incredibly difficult for patients to set up at home and can be very confusing . Robotic physical therapy systems also usually deal with a single joint to simplify their control schemes . This means most RAT TST devices only utilize specific muscle groups in a specific part of the body. However, raspberry plant container most people in stroke recovery are attempting to strengthen neural pathways in entire limbs, so RAT for multiple muscle groups can take up a huge amount of space and too much money for at-home usage throughout all of recovery to be feasible. All of these considerations have led to a market where only physical therapy centers and the few who can afford to buy, store, and replace such customized devices are able to utilize them.Drawing therapy is effective because it requires the coordinated and deliberate use of several muscle groups together. In fact, the steadiness, active range of motion, and spatial awareness required to draw a circle have led to the process being used as a common metric of stroke rehabilitation . Increased circle size can be linked to increased active range of motion, as multiple muscle groups must be fully engaged to create a large shape . The roundness of the circle is connected to both hand stability and coordination of muscle groups, as uncoordinated muscles lead to very eccentric ellipses . Creating a robot that can automate this test would allow assessments to be conducted at home or in group settings, reducing the frequency of physical therapy sessions, allowing check-ins to be conducted online, and creating more data for physical therapists to monitor recovery between sessions.

All of these effects would greatly reduce the costs of physical therapy for patients and create better quality remote health-care. Drawing therapy can also help with the psychological issues that can come with a stroke by aiding patients in creative expression and teaching new skills . Many stroke patients are frustrated with their sudden loss of fine motor control, so providing an avenue to gradually regain control where patients can see visible progress helps ease their concerns. Stroke patients often feel isolated and have identity issues due to their loss of motor function and difficulty with the coordination necessary to engage in old hobbies . Drawing and other forms of expression in a group setting can help mitigate these symptoms by providing community and helping the patient develop new hobbies. These hobbies in turn can help motor function as the repetitive and precise motor movements necessary to create art can improve coordination and aROM. Additionally, many artists and professionals who want to draw steady straight lines or curves without using software in a digital medium may benefit from a device that would aid in this task.There are no commercially available assistive devices designed to aid in drawing therapy due to the complexity of predicting both motion and the physical location of the hand to ensure images are drawn as intended, however, it has been proposed that this could be resolved by utilizing surface electromyography sensing, internal measurement units, or similar sensors . Similarly, there are no commercially available robots designed to collect data during circle drawing tests, a common way to determine the active range of motion and recovery of stroke patients . Through our complex controls scheme, we hope to continue to develop a robot with a high enough prediction accuracy, at a marketable cost, to eventually create a robot with the potential of aiding physical therapy centers, stroke rehabilitation patients, professional artists, hobbyists with unsteady hands, and professional craftsmen. We propose a design for a robot that combines sEMG sensing, internal measurement units, and drawing therapy techniques to assist in physical therapy of stroke patients with upper limb weakness. This robotic physical therapy will strengthen neural pathways for motion of the entire arm, providing a way for circle drawing tests to be accurately conducted and drawing therapy to be less frustrating at home. Our design aims to be compact, affordable, and accessible for at-home usage, allowing users to draw with assisted-as-needed technology that will incorporate feedback on their task-specific training , while they attempt to improve their active range of motion or quantify their recovery process.In 2021, we created a series of Arduino robot prototypes as proof of concept for this idea. As part of a course, these initial prototypes were created under extreme budget and time constraints. However, this also means that a final product will likely be affordable for the average consumer; we calculated that our final prototype’s components only totalled a cost of $69, with the potential for a total of $34.39 each in small batch production. Arduino prototypes are on display in Figure 1.Due to the initial budget constraints, our initial prototypes were created with Arduino components and 3D printed parts, as we had access to these from prior coursework. The robot consisted of three main sections: the circuit board, the servo motors, and the 3D printed arm. Our initial circuit board was an Arduino Uno, chosen due to its versatility with hobbyist parts, its utilization of the C++ programming language which we were familiar with, and the fact that we already had access to it. We 3D printed a case for the board to protect it and act as a weight to ensure users could not easily knock the arm over. The purpose of this component was to store code, process the input from our sensor scheme, and anchor the rest of the robot. In addition, we added an internal power supply to this prototype to increase portability. The servo motors are the method through which the robot stabilizes the user’s motion. We used two metal sg90 servos with analog feedback, as this servo model is very cheap, hardy, and can read position while allowing for a large range of motion.

The Kruskal–Wallis analysis of variance, a non-parametric equivalent of ANOVA, was used to compare groups across species. When significant, this test was followed by pairwise Mann–Whitney U-tests, and the p-values were corrected for multiple comparisons using the Bonferroni technique. In this and subsequent analyses, the Bonferroni adjustment multiplies the p-values we obtained by the number of comparisons performed . For the morphological comparisons, we tested the following models: number of bristles ¼ f and ovipositor area ¼ f. The Kruskal –Wallis analysis of variance was used to compare groups across strains. When significant, this test was followed by pairwise Mann–Whitney U-tests, and the p-values were corrected for multiple comparisons using the Bonferroni technique.We sought to determine the susceptibility of four types of fruit to each of the four species of flies. They included a fruit with relatively soft skin and fruits with much tougher skin . Our aim was to discover whether females of each species were capable of penetrating the intact skin of ripening fruit. We were aware, however, that the failure of females of a given species to lay eggs in a fruit could be a sign of either their inability to puncture the skin or a general aversion to fruit of that variety.

To distinguish between these possibilities, we counted separately the number of eggs in the exposed region of the fruit, blueberry plant pot which included the pit and any regions where the skin had naturally broken, and the undamaged fruit body. The absence of eggs in both the easily accessible exposed region and the unexposed area would indicate an aversion to the fruit, while the presence of eggs in the exposed area, and their absence wherever the skin was intact, would be consistent with the hypothesis that the skin was acting as a barrier. We found that all four species of flies laid eggs in the exposed portion of the four fruit varieties , indicating that none of the species had a strong aversion to using these fruit as larval food sources. The eggs that were laid in this region were found in a number of positions. The egg was sometimes embedded in the fruit with the respiratory filaments pointing outwards, but it could also be found in the opposite orientation or lying sideways . In the raspberry assay, a significantly larger number of eggs were laid in the exposed region by D. mimetica and D. biarmipes females when compared with D. suzukii and D. subpulchrella . This may be a consequence of the fact that the latter two species appear to have had no difficulty puncturing intact raspberry skin , and therefore did not gravitate to the far smaller exposed area in the pit.

These results contrast with the data from both varieties of grapes, where all four species laid eggs in the exposed portion, but D. suzukii laid over three times as many as the other three species. In the case of cherries, both In the raspberry and cherry assays, only D. suzukii and D. subpulchrella laid eggs in areas of the fruit where the skin was intact . There was no significant difference between these two species in the number of eggs laid in either fruit. We also found cases of punctures that lacked egg filaments , a possible indication that an egg-laying attempt was unsuccessful. These were relatively rare in both the raspberry and cherry assays , suggesting that neither species had much difficulty laying their eggs in these fruits. The finding that D. subpulchrella flies are capable of puncturing the skin of cherries and raspberries raises the question of whether this species, like D. suzukii, could be a threat to soft fruit industries.We therefore endeavoured to determine whether D. subpulchrella larvae could survive and develop in the pulp of these fruits. For the cherry assays, we observed an average of 19.9 living larvae in each fruit 5 days after the assays were set up . For the raspberry assays, we observed the bottles 14 days after they were set up and found that an average of seven adult flies had emerged from pupae. This indicates that larvae can survive in both of these fruits and, at least in raspberries, can develop to adulthood. No punctures of any form were found in the intact regions of either raspberries or cherries exposed to D. mimetica and D. biarmipes. In conjunction with the results of §3a, the findings are consistent with the hypothesis that these species rarely or never puncture intact skin even in soft-skinned fruits and that the propensity to do so evolved in the ancestor of D. suzukii and D. subpulchrella .

The results of the two grape assays differed strongly from the other fruits. Only D. suzukii laid eggs through the intact skin of any grapes, but these were relatively rare in the Thompson grapes, and entirely absent from the red grapes. However, grapes exposed to D. suzukii showed numerous punctures without the telltale sign of egg filaments . The egg-free punctures were often found in clusters.We determined, in a separate experiment , that punctures without protruding egg filaments were considerably smaller than those with filaments . These findings are consistent with the hypothesis that D. suzukii females tried to lay their eggs in the grape bodies and often succeeded in making small holes in the skin, but in most cases failed to insert an egg. However, from a pest-management perspective, it is important to note that punctures without filaments can be sites of secondary infections. By contrast, the intact body regions of grapes exposed to flies of other species never showed punctures. However, in a few cases, we found evidence of ‘slashes’ in the bodies of grapes in bottles with either D. subpulchrella or D. suzukii . The slashes, which were not found in control fruits that were not exposed to these flies, may represent unsuccessful attempts to puncture the skin. Our results demonstrate that while all species have an affinity for grapes and will lay their eggs in exposed regions of the fruit, they rarely penetrate the intact skin of the two varieties we tested. The evidence of egg-laying attempts, plastic gardening pots especially in D. suzukii, suggests that a property of the skin may make the insertion of eggs difficult. The D. suzukii findings are consistent with the results of previous research.The fruit susceptibility experiments suggest that the ability to puncture fruit skin first evolved in, or at least was strongly enhanced in, the ancestor of D. suzukii and D. subpulchrella . We tested whether this capacity was accompanied by concomitant phenotypic changes in the structure of the ovipositor. We separated, mounted and imaged ovipositor plates from two strains of each of the four species and compared them on a number of traits. One categorical distinction between ovipositors in the D. suzukii and D. subpulchrella clade, on the one hand, and D. mimetica and D. biarmipes, on the other, is the presence of thick, pigmented bristles on the former, explaining why this type of ovipositor has been referred to as ‘serrated’ . These modified bristles are found close to the distal tip of the ovipositor , which comes into contact with fruit, while the more proximal bristles are unmodified and resemble the homologous bristles present on the ovipositors of other species, referred to in the literature as ‘thorn bristles’.

The evolution of modified bristles in this clade was accompanied by a significant increase in the overall number of thorn bristles . Interestingly, the three to four marginal bristles of the D. subpulchrella ovipositor tip are generally of the thick, pigmented variety, whereas those in the equivalent region of the D. suzukii plate are almost always unmodified . However, on the lateral side of the ovipositor, D. suzukii has more modified bristles than D. subpulchrella, and the total number of modified bristles does not differ significantly between the two species . The evolution of the serrated ovipositor was accompanied by a considerable increase in the ovipositor size , with the two species with the serrated variety having ovipositor plates with approximately three to four times the area of the two other species’ plates. By contrast, comparing the same species on wing area, often used as a proxy for body size, shows an increase of only 1.6-fold to twofold in D. suzukii and D. subpulchrella . Therefore, while there has been an overall increase in body size in D. suzukii andD. subpulchrella, there has been a disproportionate increase in ovipositor area. In order to have a quantitative method of distinguishing ‘sharp’ ovipositors from ‘blunt’ ones, we measured the length towidth ratio .While this ratio does not vary significantly among D. mimetica, D. biarmipes and D. subpulchrella , it increased markedly in D. suzukii, giving the egg-laying organ a pointed, streamlined shape . We used EFA to compare the ovipositor shape among the four species, focusing on the distal half of the structure , because this is the portion of the egg-laying organ that comes into contact with the fruit during an attempted puncture. EFA produced a series of four Fourier coefficients for each of the 25 harmonics used in the analysis. Using the PCA, we calculated PCs for each of the ovipositors . The vast majority of the variation was captured by the first two PCs, which are plotted in figure 5a. Three groupings are apparent in this figure, each indicated with an ellipse: the D. mimetica and D. biarmipes strains; the D. subpulchrella strains; and the D. suzukii strains. To help to visualize what the individual PCs represent, we used the inverse Fourier transform to reconstruct ovipositor outlines on the basis of each PC . The outlines suggest that the first PC represents the difference between a sharp and blunt ovipositor. Indeed, we found that there is a strong negative Pearson product correlation between the first PC and length to width ratio. Therefore, the horizontal axis of figure 5a shows that D. biarmipes and D. mimetica, the two species that did not puncture the intact skin of any of the fruits tested, both have relatively blunt ovipositors and do not differ significantly on this variable . Drosophila subpulchrella has a sharper distal ovipositor, whereas the ovipositor of D. suzukii is the sharpest of all. A positive value for the second PC is indicative of an ovipositor with a bulb at the tip, as is seen in D. subpulchrella . Not surprisingly, there is a strong, significant difference between the D. subpulchrella strains and those of the other three species on this PC , while the latter do not vary significantly among each other on this variable . The third PC , which only explains about 4% of the variation , appears reflective of the asymmetry in the direction in which the tip is pointed. The remaining PCs each accounted for less than 1% of the variation, and the use of an omnibus Kruskal–Wallis test found no evidence of significant differences between strains. We present a model for the evolution of a modified ovipositor in figure 6 . The appearance of enlarged bristles was accompanied by a change in the shape of the ovipositor tip, which became significantly sharper. Following this event, in one species, D. subpulchrella, a distal bulb appeared, while in the second species, D. suzukii, the ovipositor became even sharper and increasingly streamlined. The latter is associated with the ability of D. suzukii to puncture the skin of the grape varieties we assayed, which was not shared by the other species in this study.While there were numerous Japanese studies of D. suzukii infestations during the first half of the twentieth century , these reports were not widely disseminated among Western researchers. When this species was first collected in California in 2008, it was not readily identified and was initially mistaken for D. biarmipes. At the time, the only known drosophilid pest in the Western Hemisphere was the distantly related Zaprionus indianus, which had first been reported in North America in 2006, 7 years after its identification in Brazil . Although Z. indianus is a formidable threat to agriculture , it can only colonize exposed fruit. Other distantly related species, however, such asthe leaf-mining fly Scaptomyza flava, possess serrated ovipositors. In this species, the ovipositor is used to puncture a leaf, and the female then feeds on its contents.

The sensitivity of B. cinerea conidia to SO2 increases two- to fourfold for every 10°C increment between 0 and 32°C, because of the effect of temperature on SO2 absorption on the fruit, fungi and surrounding packaging . SO2 has mainly been used to control gray mold disease of table grapes, but the low application rates applied as part of the total utilization technique have not prevented the emergence and spread of brown spot during cold storage . The objectives of our two-part study were to evaluate, in vitro, the effect of temperature and SO2 concentration over time applications on fungal colony growth of three Cladosporium species , and to determine the efficacy of different SO2 concentrations over time in inhibiting the growth of Cladosporium species on artificially inoculated Redglobe berries.Cladosporium species were isolated from brown spot–symptomatic berries grown in Delano, California, from 2013 to 2015. Species isolates were obtained from symptomatic tissue placed on 150 by 15 mm petri plates with 3% rose bengal potato dextrose agar amended with 500 ppm tetracycline and 300 ppm streptomycin. Cladosporium colonies were hyphal tipped after 7 days to produce pure cultures for species identification. PCR amplification of the actin gene using primers ACT-512 and ACT- 783 and DNA sequencing confirmed the identity of species .

Species isolates were maintained in Dr. W. Douglas Gubler’s laboratory, Department of Plant Pathology, UC Davis, square pot as of June 2016. Pathogenicity was demonstrated when species isolates were inoculated and re-isolated from lesions on Crimson Seedless berries following standard protocols . Asymptomatic, nonwounded berries were removed from clusters with the pedicel still intact by clipping with sterile scissors. Berries were then vigorously washed in a 0.5% potassium chloride and 0.1% Tween 20 solution to remove surface debris. Berries were surface disinfected first in a 70% ethanol solution for 30 seconds, then in a 10% bleach solution for 5 minutes, and dried in a sterile laminar flow hood . Dry berries were aseptically distributed in triplicate on sterilized polyethylene chambers on sterilized polyethylene grids at 2°C with high relative humidity . RH was obtained with paper towels moistened with sterile deionized water and placed below a plastic grid. RH was measured with a humidity sensor .To determine the effect of temperature and SO2 on fungal growth, potato dextrose agar plates were inoculated with one of three species: C. ramotenellum, C. cladosporioides or C. limoniforme. Inoculum was grown on 2% PDA at 23°C for 5 to 7 days. Plugs of colonized PDA obtained with a sterile 4-mm cork borer were placed on a 2% PDA plate, spore side down. The baseline for all measurements was 4 mm. Inoculated plates were incubated in a polyethylene chamber at 23°C for 13 hours. The petri plates were moved to a fumigation chamber and exposed to the three SO2 treatments while still lidded.

SO2 concentrations were measured using passive colorimetric dosimeter tubes and a portable SO2 detector that continuously measured SO2 concentration inside the fumigation chambers . Dosimeter tubes were taped to chamber walls opposite the SO2 flow as well as inside of a petri plate. Untreated controls were inoculated as previously described, received no SO2 exposure but were treated and incubated identically. For each species, three petri plates were inoculated and treated in triplicate and were incubated at 2°C and −2°C for up to 32 days after treatment. Radial measurements of fungal colonies were taken to assess the effect of SO2 concentration over time on mycelial growth. Colony measurements were taken from the reverse side of the lidded petri plate. Using a polyethylene ruler, two perpendicular diameter measurements were made and averaged to determine overall colony size. The diameter of the mycelium plug inoculum was included as the minimal measurement for all treatments. Measurements were taken every 6 to 8 days during incubation at 2°C and −2°C.Inoculated berries were not wounded prior to inoculation. Inoculum was prepared with 14-day-old cultures grown on PDA. Spores were suspended in 0.5% potassium chloride and 0.1% Tween 20 in sterile distilled water. Spore density was determined using a hemocytometer, and the suspension was adjusted to 1 × 107 spores/mL by the addition of sterile deionized water. The shoulder of each berry was inoculated by placing 10 μL of spore suspension within a 4-mm Vaseline ring, to prevent the inoculum from moving. Berries were positioned with the inoculated shoulder facing up. Untreated controls consisting of berries inoculated with sterile deionized water were included for each temperature and SO2 treatment. Berry lesions on unwounded berries were measured after 28 days of storage at 2°C.Twenty-four hours after berries were inoculated with one of the three Cladosporium species, they were exposed to three concentration-over-time treatments of gaseous SO2: 100 ppm-h , 200 ppm-h or 400 ppm-h . SO2 concentrations were measured as previously mentioned. Dosimeter tubes were taped to chamber walls opposite the SO2 flow. For both studies, fumigation chambers were Sterilite containers modified to allow for a rubber tube to flow SO2 into the chamber.

Chambers were placed inside a biological safety cabinet during treatment. Chamber lids were opened once a treatment was concluded and the inoculated berries on a grid/mycelial plugs on petri plates were placed into polyethylene chamber to be stored in cold storage.Radial growth of Cladosporium colonies was analyzed using a linear mixed model approach for all three species using the R package lme4 . The mixed model used a Gaussian error distribution and consisted of SO2 treatments, temperature and species as fixed effects and replicate as a random effect. We calculated the estimated marginal means and computed all pairwise comparisons using Tukey’s honestly significant difference test. For the berry study, we analyzed the proportion of infected berries with a linear mixed model approach using a Gaussian error distribution for all species, with SO2 treatments and species as fixed effects and replicate as a random effect and computed all pairwise comparisons using Tukey’s HSD test.Radial colony growth of the three Cladosporium species on petri plates with PDA was significantly reduced by the 400 ppm-h SO2 treatment, as seen in figures 2 and 3. The 400 ppm-h concentration was most effective against C. cladosporioides when petri plates were incubated at −2°C, which resulted in no radial growth. The 400 ppm-h was also effective in slowing C. cladosporioides colony growth at 2°C; radial growth grew only from 4 to 9 mm on average. Radial colony growth of C. limoniforme and C. ramotenellum was also slowed down at 400 ppm-h,growing from 4 to 5 mm and 4 to 6 mm, respectively, at −2°C. However, the same treatment was less effective at 2°C, where, on average, fungal radial growth reached 10 mm for C. limoniforme and 12 mm for C. ramotenellum after 30 days. At the lower SO2 ppm-h , there was no significant difference in radial colony growth between any of the treated species and the untreated controls by 30 days at 2°C. All species after 10 days at −2°C had reduced colony growth for all concentrations compared with those incubated at 2°C. The colony growth of C. cladosporioides at −2°C was slower than that of the other species; and it was also slower than the colony growth of C. cladosporioides incubated at 2°C, illustrating the effect of temperature on rate of growth.In the berry study, all Cladosporium species caused disease in untreated control berries: C. cladosporioides caused disease in 65% of berries, C. limoniforme in 55% berries and C. ramotenellum in 75% of berries . By contrast, SO2 treatments significantly reduced disease incidence by Cladosporium species on inoculated Redglobe berries. The 100 ppm-h treatment reduced disease incidence to less than 25% of berries for all species. The 200 ppmh treatment was the most effective, blueberries in containers in that it eliminated disease for all three Cladosporium species tested. In the 400 ppm-h treatment, there was less than 5% disease incidence in berries inoculated with C. ramotenellum and no infection occurred in berries inoculated with C. cladosporioides or C. limoniforme. Nevertheless, there was no statistically significant difference in the proportion of infected berries between the 100, 200 and 400 ppm-h SO2 treatments of the three Cladosporium species tested .

Our results demonstrate that a single application before cold storage of at least 200 ppm-h SO2 at cold storage temperatures may be an effective tool for reducing brown spot disease incidence on grape berries, although the ability of mycelium of Cladosporium species to grow on PDA growth medium in petri plates with an SO2 application was somewhat variable. We have demonstrated SO2 of 200 ppm-h and 400 ppm-h essentially prevent disease development in Redglobe berries from the three Cladosporium species tested. C. cladosporioides exhibited tolerance at 100 ppm-h, whereas 100 ppm-h was effective against C. ramotenellum and C. limoniforme. While mycelial growth on PDA was reduced with a treatment of 400 ppm-h SO2, mycelial growth was not completely eliminated. Additionally, C. ramotenellum grows faster than the other species on a petri plate with PDA, regardless of temperature or SO2 treatment. Unsurprisingly, in the berry study, inoculation of C. ramotenellum resulted in the highest disease incidence in the untreated control berries. Our study also confirmed the critical importance of maintaining table grapes below 0°C and above their freezing point to maximize their storage life potential .These results are promising for the control of brown spot during long-term storage and export of Redglobe table grapes. Although not much is known about the natural berry infection by Cladosporium, we do know that infection can occur on the surface through the epidermis and may also be a latent infection where infection starts from within the berry. Berries used in our experiments were surface sterilized before inoculation so we could be sure that infection occurred through the epidermis with our inoculum; thus, we can conclude the SO2 application prevented disease from of the Cladosporium surface inoculum. However, SO2 may not be adequate to prevent infection resulting from latent infection, causing internal rot of the berry and inoculum production, but may inhibit the resulting spread if applied after the emergence of fungal growth . This work demonstrates a potential management strategy of brown spot post-harvest. In practice, multiple factors can influence the amount of gaseous SO2 that comes into contact with the berry surface, including air flow in storage, sorption of packing materials and the amount of condensation on berries, which can absorb gasses. Additionally, a single application of SO2 is typically not as effective as multiple SO2 applications, but multiple applications may not be feasible during transport . Total utilization technique focused on gray mold control applies SO2 at 100 to 150 ppm-h, which may be too low to be effective against brown spot, especially under cold storage conditions. Future studies should be done to improve sampling and the detection of brown spot in the field prior to harvest as well as during post harvest, so this research can be applied to known compromised lots in transport and better determine the efficacy of this practice under commercial conditions.Macrotunnel production has been increasing in coastal counties of California and is poised for expansion due to its recent adoption as a standard practice by the U.S. Department of Agriculture . In high tunnel production, crops are grown within plastic-covered structures to enhance crop performance, extend production seasons and to protect crop quality. While most caneberries, some strawberries, cut flowers, herbs and leafy greens are widely grown under plastic in California, contributing $1 billion to the state’s economy, in other states small fruits, melons and nuts are also grown in high tunnel systems. This interest in plasticulture tunnels is driven by many factors: increased production due to season expansion; reduced exposure to deleterious weather events; consumer demand for fresh, local produce; and national interest in reducing transportation-related greenhouse gas emissions, amongst other concerns. Unfortunately, it comes at a time when climate-induced weather pattern changes, particularly shorter-duration, higher-frequency storm events, are expected to become the norm . The plastic covering hoop structures can reduce the available permeable surface of a field’s production area by over 90%, which increases the volume of water likely to run off a field in a storm event . During rains, water intercepted by plastic covers is channeled into post rows , accelerating soil erosion, especially on slopes, which ultimately degrades surface water quality. In California, surface water quality is regulated by the State Water Resources Control Board through the Irrigated Lands Regulatory Program .