Ghost of invasion past: legacy effects on community disassembly following eradication of an invasive ecosystem engineer

Pamela L. Reynolds
Jess Glanz
Sylvia Yang
Hann Sylvia

Abstract

By changing ecosystem processes and altering the physical landscape, invasive ecosystem engineers can have substantial impacts on ecosystem functions and human economies and may facilitate other non-native species. Eradication programs in terrestrial and aquatic systems aim to reverse the impacts of invasive species and return the system to its pre-invasion conditions. Despite an extensive focus on the impacts of both native and non-native ecosystem engineers, the consequences of removing invasive ecosystem engineers, particularly in coastal ecosystems, are largely unknown. In this study, we quantified changes in a benthic community following the eradication of the invasive ecosystem engineer, hybrid cordgrass Spartina, in San Francisco Bay, California. We used field experimental manipulations to test for persistent effects of both aboveground and belowground structural modifications of the invasive plant on the benthic community. We found significant effects of the invasive plant more than four years following eradication. Experimental modification of the above- vs. belowground structure of this ecosystem engineer revealed taxonomic specific effects resulting in hysteresis in the recovery of the benthic food webs. We found that these "legacy effects" resulted from two specific mechanisms: (1) delayed breakdown of belowground structures (stems, roots) and (2) persistence of other invasive species whose invasion was facilitated by the ecosystem engineer. Both of these mechanisms are likely to occur in similar systems where belowground structures breakdown more slowly or where other associated long-lived invaders persist. Our work is among the first to quantify the slow rate of change in food web and community processes and the persistent legacy effects of an invasive ecosystem engineer in a coastal ecosystem. We suggest that this delayed transition to pre-invasion conditions could resemble an alternate state that would be misidentified without a sufficient monitoring interval or recovery duration, with consequences for future management and restoration activity planning.