Soil Metabarcoding Helps Identify Recalcitrant Taxa From Chaparral Seed Banks

Aims Soil seed banks serve as reservoirs of plant diversity, but traditional germination‐based methods often fail to detect dormant or recalcitrant species, limiting their utility in ecological studies. Metabarcoding is a promising method to supplement traditional approaches to characterize seed bank composition, yet its adoption has been hindered by a lack of validation against traditional methods. This study aimed to (1) compare metabarcoding protocols varying in nucleotide type, barcode region, and sample preparation, and (2) test how well metabarcoding captures seed bank diversity in a degraded chaparral ecosystem. Methods We conducted controlled experiments to evaluate DNA and RNA extraction, amplification, and sequencing success using three common plant barcodes (ITS2, rbcL, trnL) applied to seeds and plant tissues embedded in sterilized sand or field soils. We also tested whether germination stimulants (water or gibberellic acid) influenced sequencing yield. We then applied metabarcoding protocols alongside traditional germination assays to homogenized soil from five chaparral sites, identifying taxa using rbcL and ITS2. Location Chaparral in Los Padres National Forest, Ventura County, California, USA. Results DNA consistently outperformed RNA in extraction reliability, sequencing success, and taxonomic breadth. Water treatment altered community composition without significantly hindering sequencing success rates. Across chaparral samples, ITS2 and rbcL identified similar numbers of taxa at genus or super‐genus levels but showed limited species‐level resolution compared to germination assays. Metabarcoding detected more genera from smaller soil volumes than germination assays. Taxa that are typically challenging to germinate, such as fire‐followers and hemiparasitic plants, were among those identified by metabarcoding and not by traditional methods. Conclusion Integrating seed bank metabarcoding with traditional germination assays improves the completeness and efficiency of plant community surveys in degraded, disturbance‐adapted ecosystems. DNA‐based metabarcoding with rbcL and ITS2 enhances detection of elusive or recalcitrant taxa and can strengthen restoration planning when paired with ecological knowledge.