Recipient Organization
SOLARID AR, LLC.
267 FAYES FOREST RD
CLINTON,AR 720316012
Performing Department
(N/A)
Non Technical Summary
As global populations reach unsustainable levels, farmers must produce 70% more food in less than 50-years. Governments are allocating resources and financial markets are increasing investments into improving agricultural processes at record levels. The agricultural industry is consolidating, acreage per farm is increasing and profit margin decreasing where farmers are searching for operational efficiencies to maintain profitability. The expense of growing crops (almonds) is divided into two types - fixed and cultivation expenses. Fixed expenses include land, trees and equipment that average 50% of this total. Cultivation expenses include 1) irrigation, 2) herbicides, 3) fertilizers and 4) pesticides that change annually but can be improved by consistent monitoring and responding with precise applications. The first three cultivation expenses are normally managed by routine, standardized schedules and are obvious and normally not difficult to control. For example, irrigation is often supplied by sub-surface micro-sprinklers automated by timers and fertilizers are added into the water system on an annual schedule.However, for pesticides, most damaging insects are nocturnal, and type of timing of infestations and densities are difficult to realize and expensive to monitor. There is a narrow threshold between 2%-to-4% of crops damaged by pests that can determine profitability. The average labor expense to monitor insects is $40/acre and cost of pesticides in orchards is $118/acre, which together comprise almost 50% of cultivation expense, making their reduced use a priority for the orchard industry. Automated identification of damaging species has been an industry objective for over 50 years. In the next five years, the orchard industry will increase by 25% and use of products like SolaRid will be in high demand to reduce insecticide volumes.This research will demonstrate the effectiveness of artificial intelligence to automate identification of damaging species of insects to reduce crop losses and quantify densities to determine optimum timing of responses to standardize pesticide applications reducing 1-to-3 insecticide applications ($50/acre) per season. This increases crop production and reduces operating (cultural) expense by as much as $18,000/acre while also limiting farm employee exposure to toxic insecticides.
Animal Health Component
33%
Research Effort Categories
Basic
34%
Applied
33%
Developmental
33%
Goals / Objectives
Because farmers must produce 70% more food, themajor goal of theresearchin our awarded USDA grant # 13496383 is to generate empirical evidence of the effectiveness of artificial intelligence technology used in a Precision Agriculture product, SolaRid. Experts state that if 15% of farms adopt precision agricultural techniques, crop yield would increase by 15% while reducing greenhouse gas emissions and inputs by as much as 20%. Analysts estimate that the Precision Agriculture market will reach $22 billion by 2025, growing at a CAGR of 9.8%. SolaRid is used to automate identification of species of insect to reduce the $47 billion of crops destroyed by pests and to quantify infestation levels for timely responses to reduce the $25 billon farmers spend on chemical insecticides. The United States spends $10 billion responding to adverse health and environmental impacts of insecticides.This research specifically targetsAmyelois transitella,aka,Navel Orange Worm (NOW), the primary pest in California almonds, which causes more than $900M of annual losses and expenses to nut crops. Insects have multiple reproduction cycles with exponential growth; early detection of initial infestations and precise responses can significantly reduce losses in crops and volumes of insecticides applied, improving health. Navel orange worm currently has four reproduction cycles, (generations) each year of which each reproduction population increases by 1900%, and the number of generations are increasing to five due to global warming.There are five measurable objectives of the USDA research.1) Can the artificial intelligence system operate automatically in an outdoor environment?2) Does Insect Control Device report activity and population phenology of the navel orange worm (Amyelois transitella)?3) Is this approach consistent or superior to the existing monitoring systems to optimize timing of pesticide responses?4) Are more females captured with light attractants?5) Can the SolaRid system efficiently distinguish male and female NOW?Theobjective of the work planis to evaluate the artificial intelligence technology (SolaRid) against conventional monitoring products. The SolaRid system using pheromone and female bait attractants is compared to these attractants, non-automated sticky traps; and its light attractants, compared to these traps (questions 1-3). In addition, the efficiency to capture female species using light attractants is compared to baits, (commercially sourced pistachio mummy bags, or 30% sweet almond oil in almond meal) in sticky traps to evaluate the potential of an attract-and-kill approach (question 4). Finally, data collected provides verification of the ability of SolaRid to distinguish male and female navel orangeworm, (question 5).This research leverages the expertise and resources of the teams. Dr. Rijal is the principal investigator for the UC ANR team, responsible for the protocols/supervision of staff members conducting in-field trials. Dr. Burks, USDA ARS is responsible for in-lab work, analyzing results of the AI technology. Mike Strmiska, a licensed Pest Control Advisor, will monitor field trials and work with PCAs to develop procedures. Randy Sasaki will overseeresearch and interoperations between the teams.
Project Methods
Work PlanA 16-week trial will be conducted at between June 20 and October 3, 2022. One location of the trial will be in StanislausCounty (plot1/Student1), and the other in Fresno County (plot2/Student2). With only two units available, the experiments will be conducted using replicates in time. The replicate blocks will be three-week periods. Treatments will be the type of attractant (pheromone, female bait, or light). Treatments will be assigned randomly within the three-week replicate blocks, with separate randomization for the two locations. Plots will consist of a SolaRid unit on the edge of a row, with multiple pheromone and bait traps extending in both directions.InJune, (start date), Students will activate light attractants and insert pheromones (replaced every 4-6 weeks) and bait attractants (replaced every 12 weeks) into SolaRid in bothtest orchard locations and begin pest monitoring(8hrs/month), on a form provided by the Company. At the same time, staff members will hang a total of 6 pheromone-baited wing traps, with 3 pheromone-baited wing traps extending in each direction from the central SolaRid units. More bait traps than pheromone traps are used because capture is usually lower on bait traps, and the bait traps are at closer intervals because, unlike pheromone traps, there is no evidence of mutual interference for bait traps. The pheromone will be a mega-lure, which attracts navel orange worm with equal efficiency regardless of the presence or absence of mating disruption. Presuming 22-foot row spacing, these traps will extend a total of 792 feet, and thus can be accommodated on the edge of a square 40 acre orchard (1320 feet on each side). There will also be a total of 10 traps baited with 30% sweet almond oil in almond meal, with five traps in five adjacent rows between the SolaRid unit and the first pheromone trap on each side. Staff members will monitor/record, bi-weekly (16hrs/month), during the in-field trials, target species counts. Their service of traps at the start of trials coincides with the second flight of NOW (June - July), when ambient temperatures are consistently high (85° - 95° F). Dr. Rijal and PCA Strmiska will maintain communication with orchard owners throughout the trials and confirm the start date is afterinsecticide applications and document date, type, and volume (per acre) of applicationsthereafter in both orchards. SolaRid's' DDT algorithm will be reset to zero by staff members at the beginning of each generation.Twice per month, S1/S2 will attach a bag on SolaRid that collects killed insects, and return 72 hrs later, photograph insects collected, count/record NOW species, email photos to Drs. Rijal/Burks and the Company, bag, label and freeze contents for analysis by Dr. Burks, as needed. Once per month, S1/S2 will set-up a camera provided by the Company (2hrs/ month) and videotape the AI - detection zone operating from 6:00 PM - 5:00AM PT and email video recording to the Company, to be used for the final report. The data from both orchards will be referenced by Drs. Rijal/Burks (10hr/month), and Sasaki/Strmiska each month to SolaRid's accumulated data of ambient temperatures, IDI, and AI identification to analyze/record anomalies and/or consistency between the two orchards through this and the third (August) and fourth (October) reproduction cycles.The technical objectives and outcomes of this research will enable the formation of a comprehensive Precision Agriculture monitoring and insect control device (ICD). The following questions to be answered in order to generate empirical evidence of the SolaRid monitoring system efficacy will be completed in 6-month field trials.1) Can the artificial intelligence system operate automatically in an outdoor environment? 2) Does ICD report activity and population phenology of the navel orange worm (Amyelois transitella)? 3) Is this approach consistent or superior to the existing monitoring systems to optimize timing of pesticide responses?4) Are more females captured with light attractants? 5) Canthe SolaRid system efficiently distinguish male and female NOW?Questions 1-3will be tested using an appropriate model within the generalized linear modeling framework. We will apply a simple 2-way ANOVA, with attractant and replicate in time as factors. Question 4 will be tested by comparing the total capture, across all five weeks of navel orange worm with the SolaRid units with the total capture of all wing traps with the comparator attractant that week. For example, SolaRid's light attractants, controlled from user dashboards, will operate on weeks 1, 4, 7, 10, and 13. To testQuestion 4, all NOW detected in the SolaRid unit during those weeks will be compared to the mean of the sum capture in the 10 female traps over those 5 weeks, using a 95% confidence interval obtained using the Lab assistant or Welch t-test. If the SolaRid's light attractants captures fewer or more moths than the 95% confidence interval, then this is evidence that one of the trap types captures more efficiently than the other. We hypothesize that the SolaRid system, when using physical (light) attractants, captures more females than traps baited with female baits (almond oil/meal). If light is indeed more effective for females, then this is an important advancement in monitoring as available data states that female trapping is more predictive of damage/losses. This predictive data is the basis for the development of an improved attract-and-kill, mass-trapping device.The test ofQuestion 5will presumably require recovery in a collection bag of all NOW entering the SolarRid trap in the weeks that lights are used as attractants. The sex ratio of NOW as determined by the artificial intelligence can be compared to the sex ratio as observed in moths collected in bags during the same time period. This analysis will be conducted using 2 x 2 contingency with chi-square statistics, or a generalized linear mixed model with a binomial distribution. The final report will be reviewed and edited by Drs. Rijal/Burks and the Company for approval; and used to create procedures for the BOS.In August, mid-point of research, Sasaki will return to CA. There, he will review research data with Drs. Rijal/Burks and work with the Almond Board of CA explaining AI technology developments and associated research activities and begin market assessment, conducted by ABC. Results of these activities are intended to create actionable intelligence for data driven management decisions, developed during this phase I research.The timing of deployment, stage of almond growth, SolaRid system settings, ambient temperatures, insect densities, biofix events detected and insecticide responses will be monitored, recorded, and videotaped (as appropriate) by UC ANR staff members retained in this research to be used in a final report and to create procedures for the Business Operating System (BOS) documentation. This will include what are "good agricultural practices" and procedures for the application of SolaRid that result in actionable intelligence for producers to improve management of abiotic and biotic stresses.