Long-term persistence of summer annuals in soil seed banks of seasonally dewatered mudflats

In annually flooded and dewatered mudflats, does the big flush of germination following dewatering exhaust the seed bank, and if not, how long can seeds remain viable and nongerminated under annual flooding cycles? Soil samples from mudflats in Land Between the Lakes in western Kentucky and northwestern central Tennessee (USA) were collected in 1990, 1991, and 1992 and annually flooded in winter and kept moist during summer for 15, 14, and 13 years, respectively. Beginning in 1996, all samples were stirred when dewatered in June. Seedlings were counted and removed, thus preventing any input of new seeds. During the study, seeds of 37 species germinated, and there were 2 to 8588 seedlings m−2, depending on the species. Ten of the species germinated in only one set of samples (2–4 seedlings m−2) in 1 year, but the number of years before germination ranged from 1 to 9; one species germinated in two sets of samples (2–4 seedlings m−2) after 6 and 9 years. Eighteen species germinated in 2–12 of the years but not necessarily in consecutive years. Eight species germinated in 14 and/or 15 years. Regardless of seed bank size, seeds of many species lived in the samples for extended periods of time with 31 (83.8%) germinating after ≥ 5 years and 19 (48.6%) after ≥ 10 years. New species appeared in the 1990, 1991, and 1992 samples after 8, 9, and 11 years, respectively. Thus, if the seedling emergence method is used to study mudflat soil seed banks and monitoring is terminated after only a relative short period of time, the size and species richness of the seed bank may be greatly underestimated.