The ups and downs of dormancy: the importance of accounting for prolonged dormancy in plant populations

Background/Question/Methods

Many perennial herbs can forgo emergence and spend one or more years belowground before appearing aboveground again.  Accounting for this prolonged dormancy (PD) could be essential for estimating plant abundance for management and conservation.  If dormancy rates vary between years, surveys could under-estimate abundance or over-estimate the effective population size.  If dormancy is common or lasts for multiple years PD could also alter the results of population projections and population viability analyses (PVA). To address these questions we evaluated PD rates and the size of the belowground population for two common North-American wildflowers for which precise estimates of dormancy rates could be made. 

Our dataset consists of individually tagged aboveground Maianthemum racemosum (hereafter MARA) and Polygonatum biflorum (hereafter POBI) followed from 2003-2010.  During this time all aboveground plants were identified and newly emerged plants marked.  We used mark-recapture models to estimate the probability that an aboveground plant would become dormant the next year.  The size of the belowground population each year was calculated by tallying the number of plants that were not present that year but emerged in the future.  The duration of dormancy was estimated by extrapolating the distribution of dormancy times using a logarithmic function.

Results/Conclusions

At any given time between 40%-60% of the plants were belowground.  Dormancy rates varied between 0.09 (±0.05 95% CI) and 0.475±0.08 for MARA, and 14±0.04 and 0.63±0.06 for POBI.

New plants emerged from dormancy each year of the study.  The abundance of aboveground MARA increased by 25% from 2003-2008.  In 2009, however, DA abundance dropped to 5% below the initial abundance in 2003.  POBI showed similar fluctuations.  Surprisingly, we found that dormancy can potentially last up to 13 years and that  >50% of plants remain dormant for >3 consecutive years.  Preliminary investigations with matrix models indicate that removing or mischaracterizing dormancy from the model significantly alters estimation of population growth rate.  

This annual variation in abundance and discrepancy between aboveground and total abundance have important implications for plant surveys, monitoring, and population modeling.  Our study shows that dormancy can be the principal driver of plant abundance and population dynamics and that short-term surveys would mischaracterize population size.  Plants that remained dormant for >5 years contributed to the observed variation in dormancy and neglecting this portion of the population could bias population projections and PVAs.