AGRICULTURAL DETECTION DOGS: MEASURING CAPABILITY AND ENHANCING CAPACITY

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: ACTIVE
• Funding Source: AFRI COMPETITIVE GRANT
• Reporting Frequency: Annual
• Accession No.: 1025175
• Grant No.: 2021-67013-33715
• Cumulative Award Amt.: $475,000.00
• Proposal No.: 2020-06483
• Project Start Date: Feb 1, 2021
• Project End Date: Jan 31, 2025
• Grant Year: 2021
• Program Code: [A1181]- Tactical Sciences for Agricultural Biosecurity

Project Director
Hall, N. J.

Recipient Organization
TEXAS TECH UNIVERSITY
(N/A)
LUBBOCK,TX 79409

Performing Department
Animal and Food Sciences

Non Technical Summary
Disease and invasive species cause significant agricultural economic damage. Specifically, powdery mildew and the spotted lanternfly comprise over 300 million dollars in annual losses with devastation to grape crops. To reduce and mitigate these losses, real-time detection in the field is key for targeted removal of spotted lanternfly eggs and precise fungicidal application for powdery mildew. Agricultural detection dogs maybe a precise, accurate, and scalable solution. Our objective here is to scientifically evaluate the capacity and capability of detection canines as an early detection tool in the fight against invasive species and diseases. We expect that detection canines will demonstrate remarkable olfactory sensitivity for the detection of agricultural threats and that a citizen science approach will demonstrate that canine detection technology may be readily and cost-effectively deployed at scale. Our first two aims will, for the first time, measure canine detection capability using precise automated behavioral testing equipment for dogs' detection of powdery mildew and spotted lanternfly eggs. Our third aim will measure the USA capacity for a large detection dog network for agricultural targets. We will leverage our partnership with the National Association of Canine Scent Work, which has registered over 20,000 dogs in detection work, to train 70 certified dog teams to detect powdery mildew and evaluate the teams in real-world deployment tests in active vineyards. Together, we will thoroughly define canine detection capability for two biosecurity threats targeting grape vines and evaluate the USA capacity for an agricultural detection dog network.

Animal Health Component

30%

Research Effort Categories

Basic

70%

Applied

30%

Developmental

(N/A)

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
215	1139	1060	50%
712	1139	1060	50%

Knowledge Area
215 - Biological Control of Pests Affecting Plants; 712 - Protect Food from Contamination by Pathogenic Microorganisms, Parasites, and Naturally Occurring Toxins;

Subject Of Investigation
1139 - Grapes, general/other;

Field Of Science
1060 - Biology (whole systems);

Keywords

agricultural detection dogs

powdery mildew

spotted lantern fly

sustainability

Goals / Objectives
Agricultural loss due to insects and pathogens exceeds 40 billion USD a year. Ourlong-termgoal is to reduce the impact of invasive speciesandagricultural diseases through early in-the-field detection mechanisms. The goal of this project is to evaluate the capabilities of agricultural detection dogs to detect two threats to the United States' 6.5 billion dollar grape industry and evaluate the United States' capacity for rapidly deploying agricultural detection dogs to remediate and prevent spread of inavsive insects or pathogens. We will achieve this project's goal through a series of objectives: 1) Evaluate canine olfactory capability to detect and discriminate Powdery Mildey in grape plant leaves determine the earliest stage of infection detectable with canine olfaction, 2) Determine canine olfactory detection accuracy for spotted lantern fly eggs and quantify dogs' limit of detection, and 3) Assess the feasibility of training and deploying citizen science owner-dog teams as a cost-effective canine detection technology for producers.

Project Methods
Efforts:The results from this study will be disseminated through various Extension outlets, including, but not limited to: the Virginia Vineyard Association annual meetings, Virginia Cooperative Extension viticulture meetings and workshops, online videos and meetings, Virginia grape disease update blog (http://grapepathology.blogspot.com), and extension bulletin. This project will further provide hands-on research training for one post-doctoral scholar and one graduate student. Additionally, we will involve a team of 70 citizen-scientist dog owners, providing educational opportunities on the outcomes of the project. Lastly, if successful, this project will also produce the largest and most distributed network of agricultural detection dogs in the United States using an easily scalable and cost efficient model.Evaluation:Validation of dog's capability to detect target scent: Dogs (n=10) will be trained to differentiate the scent of grape powdery mildew on grape leaves (target) from healthy grape leaves. Additionally dogs will be trained to discriminate grape powdery mildew from rose powdery mildew on grape leaves (non-target), grape ripe rot pathogen on grape leaves (non-target), other non-target molds (non-target), and other non-target objects (e.g., gloves). We will use novel canine automated olfactometry to provide precise and controlled odor stimuli to canines while controlling for unintentional cuing in training.Dogs will also be trained to detect Spotted lanternfly eggs from a variety of distractors. Samples of the target and non-target fungal species will be collected from vineyards and other fields in Virginia, and clean, single-spore cultures will be grown in the appropriate medium or plant surface. Spotted lanternfly eggs will be collected from northern Virginia where this insect has been established. These plates and insects will be shipped to the training facility for the training.Milestone: Detection rate of 80% accuracy or higher (80% correct responses). We will calculate sensitivity and specificity measures for each dog, and we anticipate an overall detection accuracy of 80% [range: 70-100%].Evaluation of threshold of detection: After dogs are successfully trained to detect powdery mildew and spotted lantern fly eggs, we will, for the first time, evaluate canine detection sensitivity limits for each target. For powdery mildew, we will measure the minimum time from powdery mildew inoculation that a dog can accurately respond and detection infection. For spotted lantern fly egg, we will use air-dilution olfactometery to measure the maximum air dilution that dogs can still detect spotted lantern fly egg masses, providing a measure of the canine detection limits. Threshold detection limits will be evaluated using the psychophysical procedure of 2-down 1-up descending staircase. This adaptive procedure efficiently measures detection sensitivity limits.Milestone: Provide the first measures (and 95% confidence intervals) of the earliest stage of infection powdery mildew is detectable and the threshold detection (in parts per million of vapor saturation) for spotted lantern fly egg masses. This will allow us to develop empirical models of successful detection in applied scenarios.Evaluation of the feasibility of training citizen science owner-dog teams to detect powdery mildew: Through partnering with the National Association of Canine Scent Work (NACSW) we will recruit four cohorts of 40 owner-dog pairs. The recruited owner-dog pairs will be teams that have already demonstrated the ability to learn scent tasks and will have earned at least a Nosework 3 (NW3) title (and passed initial detection certification). We expect some attrition and estimate 20-25 owner-dog pairs per training group (80-100 pairs total) will move to field testing. We will train dogs on powdery mildew because it is found in vineyards across the U.S. (rather than being regionally specific). Training for the dogs will follow the standard protocols of NACSW with minor adjustments as needed for handling of the specific target odor. After preliminary training, dogs will be tested on an odor recognition test for powdery mildew. This will be used to determine if the dog has learned to detect powdery mildew and can discriminate the presence of it from other environmental odors. The test will follow the protocol of NACSW Odor Recognition Trials (ORT).Milestone: 70 dogs pass the ORT. To pass the ORT, the dog must indicate the location of the odor by engaging in a specified response that the owner identifies as their indication within 3 minutes of search time.Evaluation of success of training citizen science owner-dog teams to detect powdery mildew in a field setting: After passing an ORT test, dogs will continue training on the NACSW training protocol including: 1) refining the dog's indication or the explicit response the dog exhibits when the source of odor is detected; 2) stimulus control training in which the dog learns to emit the indication in the presence of odor rather than non-target odors; and 3) environmental complexity training in which dogs are trained to detect the odor in increasingly difficult hides (e.g., at new heights, with more distractor containers or scents) and increasingly distracting environments (e.g., outdoors) until they are training in a setting similar in complexity to a field setting. After field training, each dog will be tested on a field test for the trained target odor using the protocols of NACSW Level 2 Element Specialty Exterior tests. This field test will entail the dog being tested in an outdoor search area that is similar to the environment in which the target odor would normally be found (e.g., a plant nursery for PM).Milestone: 70 dogs pass the field test. To pass the field test, the dog must find four out of five hides within four minutes and not false alert more than once.Evaluation of success of trained citizen science owner-dog teams to detect powdery mildew during deployment and comparison to human observers: Dogs that pass the field test will be tested in a deployment setting with naturally occurring targets. This test will take place in a vineyard where at least one plant has been identified as having PM or has been infected by the experimenters for the sake of the deployment test. For any indication of PM on a plant that cannot be visually corroborated by a human observer, we will remove a leaf from the plant and send it to Dr. Nita's lab for PCR identification. The plant will be followed for another 2 weeks in case infection develops to ensure the validity of canine responses. Prior to the dogs searching in the deployment test, a trained human observer will be deployed in the search area. We will measure the human observer's latency to detect targets and accuracy. We will compare their results to the dogs' results from the deployment test.Milestone: Tested dogs each detect at least one occurrence of PM in the deployment setting. Compared to human observers, dogs detect more occurrences of PM and more quickly than do trained human observers.

Progress 02/01/23 to 01/31/24

Outputs
Target Audience:This year, as part of our community science aim, our team conducted 15 interviews on canine detection of invasive species for various organizations including The Today Show, Accuweather, and The Weather Channel. We have also provided presentations for the Pennsylvania Department of Agriculture Spotted Lantern Fly detection dog team. In addition, our community science aim engaged over 1,100 community members that participated in our online webinars. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?We have created IDPs (Individual development plans) for the post doc supported on this project, who recently received a tenure-track position in Animal Science. The project has also provided training for one PhD student and one MS student. How have the results been disseminated to communities of interest?This year, as part of our community science aim, our team conducted 15 interviews on canine detection of invasive species for various organizations including The Today Show, Accuweather, and The Weather Channel. We have also provided presentations for the Pennsylvania Department of Agriculture Spotted Lantern Fly detection dog team. In addition, our community science aim engaged over 1,100 community members that participated in our online webinars and 168 community members are actively participating in detection dog testing. What do you plan to do during the next reporting period to accomplish the goals?In the next year, our focus will remain Aim 1 and Aim 3, which we anticipate completing data collection within the year.

Impacts
What was accomplished under these goals? Early detection of invasives species and disease remains a critical need to reduce food security risks and agricultural loss. This project investigates novel applications of detection canines to assess their detection capability and sensitivity for (1) Powdery Mildew and (2) Spotted Lantern Fly egg masses. In addition, (3) this project investigates the capacity for citizen science approaches to broadly mobilize invasive species detection in agricultural regions in need. In this past year we have focused on Aim 1 and Aim 3 for new data collection while disseminating knowledge from Aim 2. This year, we published two additional peer-reviewed papers, leading to changes in knowledge. In these manuscripts, we have discovered (changes in knowledge) that dogs have remarkable detection capabilities for Spotted Lantern Fly egg masses with sensitivity and specificity exceeding 99%. Further, dogs can detect spotted lantern fly egg masses diluted to and air dilution of 0.9% (without an odor background) or 4.1% (with an odor background), which exceeds that of several explosives that dogs are routinely trained to detect. This indicates that dogs' sensitivity for this invasive species is as good or better than that of explosives. Lastly, we developed a method to encapsulate egg masses that prevent egg loss but provide a whole representation of the egg odor from a canine perceptual perspective. Using these training aids, we found that dogs can be used to scan wood pallets (a common site for SLF eggs) in ~10 s with over 90% sensitivity and specificity for real-world searches. We have disseminated this knowledge to canine handlers with Spotted Lantern Fly detection canine handlers, leading to changes in action. Aim 3 evaluates whether there is an underutilized capacity within the United States for agricultural detection dogs. Using a community science approach, we are evaluating whether a national network of community scientist can create a national agricultural detection dog program. This year we had over 900 interested community members participate in a webinar to learn more about Spotted Lanternfly. Later, 1,214 personnel filled out a form of interest to participate. We identified 23 trainers and 164 dog handler teams to participate in our first year of the program. Of the 164 dog teams, 88 teams passed our double-blind odor recognition test, and 48 teams have completed a field evaluation. Given that there are currently only a handful of operational SLF detection dog teams, this represents over a 10x increase in SLF detection dog capacity from a volunteer teams. Remaining teams are awaiting certification testing. Overall, citizen scientists participated from over 10 different states including: VA MD PA NC RI NY WA TX OR and CA. For Aim 1, three dogs completed olfactory training to detect Powdery Mildew (PM). Within nine training sessions all dogs were able to discriminate PM positive leaves from other distractors, including PM negative leaves, with an accuracy ≥ 80% . Furthermore, dogs were able to accurately discriminate powdery mildew from three additional plant infectious agents. Additional work is planned for the next year to complete this Aim.

Publications

• Type: Journal Articles Status: Published Year Published: 2023 Citation: Aviles-Rosa, E. O., Kane, S. A., Nita, M., Feuerbacher, E., & Hall, N. J. (2023). Olfactory threshold of dogs (Canis lupus familiaris) to cold-killed spotted lantern fly eggs. Applied Animal Behaviour Science, 261, 105880.
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2023 Citation: Nita (2023) SLF management tips, Virginia Vineyard Association, Charlottesville, VA, 18 July 2023
• Type: Journal Articles Status: Published Year Published: 2023 Citation: Kane, S. A., Aviles-Rosa, E. O., & Hall, N. J. (2023). Development and assessment of Spotted Lanternfly (Lycorma delicatula) training aids for detection canines. Applied Animal Behaviour Science, 266, 106011. https://doi.org/10.1016/j.applanim.2023.106011
• Type: Theses/Dissertations Status: Published Year Published: 2023 Citation: Cantu, A. (2023) Evaluating volatile organic compounds from the spotted lanternfly (Lycorma delicatula) using headspace odor sampling methods. Thesis: https://hdl.handle.net/2346/96380
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2023 Citation: Dickinson, S. (2023). Effectiveness of Citizen Scientist Dog Teams in Detecting Spotted Lanternfly egg masses. Association for Pet Dog Trainers. October 2023.
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2023 Citation: Nita (2023) VT Grapepathology Updates, Virginia Vineyard Association, Charlottesville, VA, 21 Feb 2023
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2023 Citation: Nita (2023) Perspectives on PD, viruses (mealybugs), and Lanternfly, Madison, VA 6 Dec 2023

Progress 02/01/22 to 01/31/23

Outputs
Target Audience:Our outreach efforts have targeted the Pennsylvania Department of Agriculture Spotted Lantern Fly detection dog team as well as the New York New Jersey trail conference. We have had meetings with both organizations and transmitted our research findings. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?We have created IDPs (Individual development plans) for the post doc supported on this project. In addition a new PhD student has started the project and will begin professional development. How have the results been disseminated to communities of interest?We have outreached to state Departments of Agriculture canine teams to disseminate our Spotted Lantern Fly results as well as conservation detection canines handlers. We have also incorporated lessons in our Animal Science classes to disseminate our specifc results as well as enhance science/research education. Two presentations on the research have also been made: Nita, M (2022) Overview of AHS AREC research, VCE Winter Conference, Online, 02/02/2022, Oral (project summary was presented) Nita, M (2022) Grape Pathology Field Trials 2022, AHS AREC Field Day, Winchester, VA, 09/16/2023, Oral (project summary was presented) What do you plan to do during the next reporting period to accomplish the goals?In the next year, our focus will turn to Aims 1 and to Aims 3. Aim 3 we anticipate to take the rest of Years 3 and Year4. We anticipate completing the two experiments for Aim 1 this year.

Impacts
What was accomplished under these goals? Early detection of invasives species and disease remains a critical need to reduce food security risks and agricultural loss. This project investigates novel applications of detection canines to assess their detection capability and sensitivity for (1) Powdery Mildew and (2) Spotted Lantern Fly. In addition, (3) this project investigates the capacity for citizen science approaches to broadly mobilize invasive species detection in agricultural regions in need. In this past year we have focused on Aim 2, having three manuscripts published and/or under review, leading to changes in knowledge. In these manuscripts, we have discovered (changes in knowledge) that dogs have remarkable detection capabilities for Spotted Lantern Fly egg masses with sensitivity and specificity exceeding 99%. Further, dogs can detect spotted lantern fly egg masses diluted to and air dilution of 0.9% (without an odor background) or 4.1% (with an odor background), which exceeds that of several explosives dogs are routinely trained to detect. This indicates that dogs' sensitivity for this invasive species is as good or better than that of explosives. Lastly, we developed a method to encapsulate egg masses that prevent egg loss but provide a whole representation of the egg odor from a canine perceptual perspective. Using these training aids, we found that dogs can be used to scan wood pallets (a common site for SLF eggs) in ~10 s with over 90% sensitivity and specificity forreal-world searches. We have disseminated this knowledge to canine handlers with Spotted Lantern Fly detection canine handlers, leading to changes in action. Aim 3 has started the recruitment phase and we have begun pilot work with our first citizen science team. We have piloted citizen science detection with two currently trained scent dogsand have met with the National Association for Canine Scent Work (NACSW) to outline their participation in the project. We are now preparing and deploying recruitment materials for citizen science participants. For Aim 1, eight dogs started olfactory training to detectPowdery Mildew (PM) positive grape leaves. As anticipated based on previous research, dogs learned the task rapidly. Within nine training sessions all dogs were able to discriminate PM positive leaves from other distractors, including PM negative leaves, with an accuracy ≥ 80% (Figure 1). This work, however, is still early and not yet ready for publication. This will be a focus for our next year's work. Nevertheless, our preliminary results indicate that dogs are proficient detecting the presence of PM using olfactory stimuli from the infected leaves. Figure 1. Dogs (N = 8) proportion of responses to Powdery Mildew (PM) positive grape leaves and to the different distractor odors present during the last training session. Dogs received a total of 20 trials where they searched three ports each one containing either the headspace of a vial containing a PM positive leaf or the headspace of a vial containing a distractor including a PM negative grape leaf. Overall, dogs' accuracy to PM positive leaf was > 85%. This indicates that dogs can identify PM positive grape leaves using olfactory stimuli emanating from the leaf.

Publications

• Type: Journal Articles Status: Published Year Published: 2023 Citation: Aviles-Rosa, E. O., Nita, M., Feuerbacher, E., and Hall, N. J. 2022. An evaluation of Spotted Lanternfly (Lycorma delicatula) detection dog training and performance. Applied Animal Behaviour Science, 105816. https://doi.org/10.1016/j.applanim.2022.105816
• Type: Journal Articles Status: Under Review Year Published: 2023 Citation: Development and assessment of Spotted Lanternfly (Lycorma delicatula) training aids for detection canines. Applied Animal Behaviour Science.
• Type: Journal Articles Status: Under Review Year Published: 2023 Citation: Olfactory threshold of dogs (Canis lupus familiaris) to cold-killed Spotted Lantern Fly eggs. Applied Animal Behaviour Science

Progress 02/01/21 to 01/31/22

Outputs
Target Audience:This past year, we reached various growers through media outlets (https://www.capitalpress.com/ag_sectors/orchards_nuts_vines/researchers-train-dogs-to-sniff-out-pests-and-pathogens/article_bd69d6c2-754b-11eb-a926-371e3a7ec78a.html ) as well as USDA APHIS detection dog handlers in which we discussed our results and findings with handlers in the field detecting the same pests. Additionally we presened at a working dog conference for scientisits and practitioners and will be presenting in March to a growers organization Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The project has provided Post Doctoral training for Edgar Aviles-Rosa in animal behavior and analytical chemistry. Training has also been provided to animal care taking staff (12 undergraduates) learning to care and provide husbandry for project animals. Additional research training has been provided to one PhD student (although not directly sponsored by the project). How have the results been disseminated to communities of interest?Results have been disseminated to an academic working dog conference (International working dog and breeding association) , have submitted to an industry conference (Penn Vet Working dog conference), held a research update and discussion call with USDA APHIS canine handlers, and have communicated our project to the press for dissemination to relevant stakeholders. What do you plan to do during the next reporting period to accomplish the goals?Over the next year, we will primarily address objective 1. Dogs will be trained for detection of powdery mildew. In addition, we anticipate publishing 2-3 papers on our acomplishments of objective 2 over the previous year.

Impacts
What was accomplished under these goals? We have acomplished objective 2; determining canine accuracy to detected spotted lantern fly and determined limits of detection. This year, we will accomplish objective 1, and the final two years will focus on objective 3.

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2021 Citation: Aviles-Rosa,E., Hall, N.J.. A detailed evaluation of Spotted Lanternfly (Lycorma delicatula) detection dog training and performance. International Working Dog Conference. October 2021.
• Type: Conference Papers and Presentations Status: Submitted Year Published: 2022 Citation: Kane, S., Aviles-Rosa,E. Hall, N.J. Development of Training Aids for Spotted Lantern Fly Detection. Penn Vet Working Dog Conference. 2022
• Type: Conference Papers and Presentations Status: Submitted Year Published: 2022 Citation: Nita, M., Updates on downy mildew, ripe rot, and canine detection of SLF eggs. Virginia Vineyards Association