Accessing limnological change and variability using fossil diatom assemblages, south-east Australia

Abstract

<jats:title>Abstract</jats:title><jats:p>Floodplain wetlands accumulate river‐borne sediments that include mixed assemblages of allochthonous and autochthonous diatoms as fossils. These assemblages have been used in river floodplain wetlands and reservoirs to quantitatively reconstruct salinity, pH and nutrients and to qualitatively infer connectivity and turbidity over periods spanning decades to millennia. High sedimentation rates in some sites have permitted sub‐annual temporal resolution; however, annual to decadal resolution is more usual. The establishment of chronologies for these sequences is often difficult owing to the substantial input of fluvially borne<jats:sup>210</jats:sup>Pb, the high spatial variability in the earliest detection of exotic pollen markers and the inaccuracy of radiocarbon approaches in dating sediments younger than 500 years. Other complexities arise from the difficulty of differentiating the influence of co‐variables in accord with the river continuum concept and identifying shifts driven by hydroseral influences independent of changes to the fluvial system. Caution is also needed in inferring lotic change from a record accumulating in lentic systems.</jats:p><jats:p>Nevertheless, substantial increases in salinity (lower Snowy, lower and middle Murray), pH (mid‐Goulburn), turbidity (upper and lower Murray and Yarra), nutrients (lower Murray and Yarra), and sedimentation rate (widespread), as well as clear shifts in trophic structure (upper Murray), have been documented for the post‐European period from regulated river wetlands across southeast Australia. A site in the lower Murray records river connectivity and water quality changes consistent with the regional Holocene climate record. Reductions in effective precipitation documented in closed lake systems are not evident in riverine plain wetlands, possibly owing to their relative complexity. The refinement of chronologies and data‐bases will allow the determination of the pre‐impact nature and variability of sites, the rates of limnological change and biological responses and the feasibility of rehabilitation targets. Copyright © 2005 John Wiley &amp; Sons, Ltd.</jats:p>