The influence of basin morphometry on the regional coherence of patterns of diatom-inferred salinity in lakes of the northern Great Plains (USA)

Sedimentary diatom profiles from saline lakes are frequently used to reconstruct lakewater salinity as an indicator of drought. However, diatom-inferred salinity (DI-salinity) reconstructions from geographically proximal sites in the Great Plains (USA) have yielded disparate results. This study explores how physical changes in lake habitat resulting from drought may affect climate inferences from salinity reconstructions. Differences in relationships among drought, lake-level change, and diatom community structure over the last century were examined for three saline lakes in the northern Great Plains with dissimilar DI-salinity records. At each site, models were developed relating available planktic:benthic (P:B) habitat area to lake-level change, and models were compared with instrumental drought records and fossil diatoms to understand how drought conditions were recorded in sedimentary diatom assemblages. The degree to which DI-salinity tracked drought variation was affected by site-specific physical characteristics that influenced the relationship between lake-level change and P:B habitat zonation within each lake. Moon Lake showed the strongest correlation between drought and DI-salinity, although this relationship was weaker during wetter conditions, as highstands resulted in a larger influx of benthic diatoms. At Coldwater Lake, a dual-basin system, P:B varied depending on lake level, which apparently reduced the correlation between DI-salinity and drought. At Lake Cochrane, the simplest and freshest of the three basins, the P:B of fossil diatoms was a better proxy for drought than DI-salinity. The integration of additional ecological characteristics into interpretations of paleoclimate records, particularly for biologically-based proxies, may improve reconstructions of regional patterns of climate variation.