Create plant reference records for model training before release

Your plant submissions arrive as single entries, causing fragmented trait records and delayed dataset preparation for models. Data stewards get consolidated reference records and consistent attributes so training runs use clean inputs before the next model training cycle.