/
Meeting 4

Attendees - Session 1

Stuart Rhea, Sarah Hovinga, Andres Ferreyra, Hiroyuki Morita, Daoud Urdu, Robert Garcia, Cale Boriack, Brent Kemp, RackWoo Kim, Marie Beuret, Clement Jonquet

Session 2

Stuart Rhea, Andres Ferreyra, Montague Dixon, Hiroko Miura, Akane Takezaki

  • Reviewed agenda, no changes made

  • Level set, recap of current model, expectations and examples

  • Reviewed use cases captured so far, last time, pointed out we are still missing use cases from food processors and CPG companies

    • Note both of the missing use cases are in the value chain where the “crop” has transitioned to a “harvested commodity”

      • Would be good to maybe draw a vertical line through the harvest node, left of that point is crop, right of that point is harvested product

    • Also need to split our manufacturers, currently considers equipment manufacturers, would also apply to other manufacturers such as seed or crop protection products

    • Question on if there should also be a regulatory aspect for use cases, agreement is important but need to figure out where to add as well as get engagement from someone invovled

      • MRL and related terminology is another area that could use some global standardization

  • Mention of 2013-15 OPPEL project by US EPA

    • Office of Pesticide Program Electronic Label pilot

    • Being used as a base for AgriGuide project in Europe since a lot of work is still relevant

    • Excel sheet added to https://aggateway.atlassian.net/l/cp/MX0ETvA4

      • Looks to be a good base for refinements and some controlled vocabularies

  • Review of EPA Active Pesticide Product Registration Information Listing (APPRIL)

    • Link to site to get the excel reviewed (new version posted every Monday)

    • Question if crop is ever different between a state list of crops and the national list (US example)

      • Need to ensure there is a means to capture geo-politically dependent context information, including different regional/national governmental levels

    • Reviewed parsed table

      • EPA Number - Column A with Reg Number

        • EPA Registration number

          • Company Number - Product Registration Number - Supplemental Distributor (Optional part, purchasing a product from another manufacturer and selling under their name)

          • Optional third section of registration umber corresponds to DP in Registration Type (Column B)

      • Site (Column T) is where the product is used, and is not agriculture specific

        • Stuart parsed the string first by comma delimited values

          • Then text in ( ) was converted to tags to convert to mostly unique values to identify dimensions of refinement

          • Have not pulled all the product to see what the tags mean, e.g. if the product can be used on tomatoes being grown hydroponically vs some use within irrigation

        • The sheet is not usable data, was just used to try and detect patterns to classify crop refinement categories

        • There are several issues with consistency

          • https://openrefine.org/ suggested as a tool to help with cleaning the list

          • e.g. Tomato vs. tomatoes

          • soil fumigation vs. greenhouse-soil fumigation may indicate there is a refinement needed for where the crop is grown greenhouse vs. open field

        • The list seems to mix the what, where, how when looking through the entries for tomatoes

          • Where - greenhouse, field(implicit), container and where on the plant foliar vs soil

          • How treatment is applied, fumigation, injected, etc.

        • Question on if there are any labels of the excel sheet that explain the status of the value

          • e.g. are any of the entries tagged as superseded, inactive, invalid, etc.

            • Does include status of the product, but not status of the tags

          • Question on what a sub-crop is

            • Color and size are likely separate things, sub-crop is more a variety

      • List is a good example of why this work is needed, very messy list that is not very machine readable

      • Note that in the Japanese MRL (Maximum Residue Limit) is partially driven by the size of the fruit, e.g. limit for cherry tomato is different than standard tomato

        • Question on if the difference is driven by actual size of the fruit (fits through a certain size sieve) or if it is the variety being of a different size

  • Canadian regulator agency is doing something similar to what we are doing, they are a little farther ahead in leveraging controlled vocabularies

  • Netherlands vocabularies review

    • Leveraged in “Closed Loop Spray” project from AgGateway Europe, list is what is used for reporting in The Netherlands

    • Field centric, but did have 5 or 6 crop refinement lists

      • Attempted to translate to English from Dutch with mixed results

      • This list does not link the sub crops to crops, at least as shared

        • No ability to show the relationships, e.g. what are all the types of tomatoes

        • There are plans underway to transform the list into a graph database or have an ontology to create these linkages

    • Curious where the list came from

    • Also brings up the need for some sort of long term maintenance of any lists this group develops

    • Many of these things change somewhat frequently so if there is not a robust mechanism to maintain the utility will drop off rapidly

      • ISO 19135 provides a framework for managing lists like this

      • AgroPortal is doing some of this already

    • The intended use column looks promising, believe there is something in the CVT from USDA that would be similar

    • eCFR Code of Federal Regulations

    • Looked at other dimensions that seems to be a good starting point for refinement categories

      • Growth environment (open air, greenhouse, …)

      • Cultivation type

      • Growth medium (soil, water, straw, container, …)

      • Crop rotation seems to be another aspect in the list

  • Homework to review and add to the catalog of vocabularies