Marine lakes of Palau barcoded specimens from transect survey with both lab identification number (M0D#) and original tube number from 2011-06-04 to 2015-07-02 (NCEI Accession 0278386)

Acquisition Description:
Sample collection:

Each lake was sampled using the point intercept transect method at no less than 10 randomly chosen sites around its perimeter (unless the small size of a lake precluded this number of non-overlapping sites). At each site, three parallel transects approximately were run 5 m apart from the intertidal (0 m) to the deepest depth accessible to SCUBA divers (i.e. the bottom of the lake, or bottom of the epilimnion, or the divers’ maximum certified depth). In lakes 8m or deeper, a line (‘the horizontal’) was placed, at eight evenly spaced target depths (1–4 m depth intervals, depending on lake), orthogonal to each of the transect lines so that small (2.0 cm diameter) cells fell over four points A–D each at 15 cm increments to the right of the transect line at the same depth. At each depth, from deepest to shallowest, the actual depth was measured in feet with a dive computer, the ‘horizontal’ was photographed from ~0.5 m distance, the substrate type was recorded, and then each cell photographed in close-up. A tissue sample of the ‘primary’ organism within each cell, i.e. the organism at the center of the cell, or if no organism in the centre then the first organisms at the periphery going clockwise from noon, or if only sediment visible, the organism within the sediment directly under the Cell was then biopsied for DNA analyses and placed in a container labeled with site, depth, and cell code A–D. Because the benthos may be three dimensional a ‘primary’ organism might also have many ‘secondary’ epibionts and/or epiphytes attached. Any organisms in the photographs but not sampled were classified as ‘tertiary’. After all four cells were sampled at a depth, the diver ascended to the next shallowest depth on his/her transect and repeated the procedure. Thus, at each randomly chosen site, we surveyed a total of 96 points from the deepest to shallowest depths of the lake habitable by macro-invertebrates and macrophytes, with two categories of exception. (1) If a lake was <8 m deep, the number of depths sampled was equivalent to the maximum depth in meters. (2) Lakes with gently sloping sides could lead to adjacent target depths being >10 m apart leading to undersampling of horizontal patchiness; in which case the transect distance between adjacent target depths was estimated and divided in half or in thirds so that no two samples were more than 10 transect meters apart. At the surface, at the end of each dive, samples were transferred to individual tubes of 95% ethanol labeled with a field number composed of lake, site, collector, depth, and cell IDs. Each evening, new samples were stored in a freezer, dive profiles were downloaded, and fieldnotes were transcribed to a standardized electronic data sheet.

Error-checking biodiversity transect files

Each evening, or as soon thereafter as possible, divers compared specimens to standardize field-identifications and all tissue samples were reconciled to the electronic data sheet for each lake using tube labels, original field notes, photographs of specimens in the field, and visual inspection of tube contents. If necessary, primary and secondary specimens were placed in individually labelled tubes of ethanol. In cases of discrepancy between electronic notes and original field notes, we edited the electronic data sheet to be consistent with original notes and corroborated this by double-checking the original photographs and inspecting tube contents. Significant changes—i.e. samples that could not be reconciled after accounting for tube transpositions, mislabeling, or mis-identification in the field—were logged in a separate file highlighting the specific change and justification. If a specimen was unable to be reconciled with notes it was discarded (this was necessary for only one specimen). Subsequently, every tissue sample was assigned an unique identifying number (M0D#) for curation; during this process, every tenth sample was double-checked for agreement between the original field number and new M0D#.