UC Davis

Natural trace element signatures have increasingly been used to track the dispersal of marine larvae, and these studies require an atlas of potential source populations with distinctive elemental signatures. To determine whether natal site atlases could be used repeatedly and to identify site characteristics that yield the best results, we built atlases in 5 consecutive years using embryos of the porcelain crab Petrolisthes cinctipes that were collected from 15 sites spanning 190 km of the open coast of northern California, USA. We analyzed the elemental composition of the embryos using a discriminant function optimization procedure to determine the suite of elements that resulted in the best reclassification success for individual sites and groups of sites each year. No single element or group of elements succeeded in discriminating the origins of embryos every year, and the reclassification success of the atlas varied at all spatial scales among years. Average reclassification success at the site level ranged annually from 39.5 to 54.3% correct, and combining sites into 2 or 3 areas improved the overall reclassification success to 72.5 to 97.7% correct. Sites with (1) distinctive elemental compositions of rocks, (2) unusual habitats, (3) consistent freshwater input, (4) consistent anthropogenic inputs, or (5) complex local oceano graphy had the highest reclassification success (up to 86.7% correct), but interannual variation in runoff reduced the temporal stability of the atlas. To improve population connectivity estimates, future trace element studies should consider these 5 key characteristics when selecting sites and anticipate temporal variation in natal site signatures. © Inter-Research 2013.