California clapper rails in San Francisco Bay: Modeling habitat relationships at multiple scales to guide habitat restoration and eradication of non-native Spartina

There are numerous factors threatening the health of tidal-marsh ecosystems in the San Francisco Estuary. The eradication of non-native Spartina alterniflora (and its hybrids with the native Spartina foliosa) is a high priority for regional scientists, due to the potential for long-term loss of structural and biological diversity if the invasion were allowed to continue. Regional non-native Spartina control efforts were initiated in 2004, led by the Coastal Conservancy’s San Francisco Estuary Invasive Spartina Project (ISP). One unexpected outcome of the non-native Spartina invasion is that the endangered California clapper rail (Rallus longirostris obsoletus) has colonized invaded marshes that would appear to be otherwise poor habitat, apparently being attracted to these sites by the unusually tall and dense plant structure associated with the non-native Spartina hybrids. This association between the non-native Spartina and the clapper rail has created a unique set of issues for non-native Spartina removal, because the benefits to the entire ecosystem of controlling the Spartina invasion need to be weighed against the localized losses of clapper rail habitat and clapper rails.

To aid in understanding this system and reducing impacts to rails, we modeled clapper rail habitat relationships using a combination of clapper rail survey data, habitat data we collected in the field, and site characteristics generated from aerial photos using Geographic Information System technology. We found a statistically significant positive relationship between a marsh’s Spartina hybrid cover and clapper rail density. We also found that rail densities were higher at marshes with higher densities of tall plant stems (50-60 cm above the marsh surface), and longer low-elevation marsh edge relative to upland edge. The most interesting finding was a statistical interaction between Spartina cover and channelization, i.e., if high values of one were present, the value of the other was not important.