Soil seed bank assay methods influence interpretation of non‐native plant management

Abstract Questions In the assessment of a soil seed bank at invaded sites, do sampling at different microsite types and use of different assessment methods yield different results, which may influence interpretation of management effectiveness and potential for native plant restoration? Location Saguaro National Park, Arizona, United States. Methods Seed bank samples were collected from two microsite types (below‐shrub canopies and interspaces between perennial plants) at plots within Pennisetum ciliare‐invaded and untreated sites, non‐invaded sites and sites with five frequencies of Pennisetum‐removal treatments. Seed bank samples were characterized by seedling emergence and seed extraction methods. Results Assay method influenced whether we detected a significant Pennisetum‐removal treatment response on Pennisetum detected in the seed bank. A treatment response was not observed using the emergence method. Using the extraction method, we found that untreated Pennisetum plots contained significantly more Pennisetum seed than removal treatments. Treatments did not affect detection of native seed banks using either method. Microsite type only had an effect on seed detected using the emergence method, where more perennial species and seed densities were detected below shrubs. Native species richness and estimated seed densities using extraction were correlated with emergence results (r = 0.52, N = 42, p < 0.001 and r = 0.56; N = 42, p < 0.001, respectively). Conclusions Results highlight several considerations for how different seed bank methods can influence interpretation of management treatments. Along with vegetation surveys, seed bank observations indicate that removal treatments successfully reduced Pennisetum in soil seed banks at invaded sites relative to invaded untreated sites. Natives were a significant proportion of seed banks, showing potential for native plants to establish after treatments of non‐native plants. Results suggest potential to calibrate seed bank methods with each other to reconcile differences. Because viability of extracted seed was uncertain, further analysis may be necessary to test which assay method best predicts future vegetation.