Multi-proxy palaeoecological responses to water-level fluctuations in three shallow Turkish lakes

Abstract

Natural or human-induced water-level fluctuations influence the structure and function of shallow lakes, especially in semi-arid to arid climate regions. In order to reliably interpret the effect of water-level changes from sedimentary remains in the absence of historical data, it is crucial to understand the variation in sedimentary proxies in relation to water level measurements. Here, we took advantage of existing water surface elevation data on three large shallow lakes in Turkey to elucidate the impact of lake-level changes on benthic-pelagic primary production over the last 50–100years. Sub-fossil cladocerans, diatoms, plant remains and pigments were investigated as biological variables; X-ray fluorescence (XRF) and loss on ignition (LOI) analyses were conducted as geochemical-physical variables on a set of 210Pb and 137Cs dated cores. Dating of the cores were robust, with the exception of uncertainties in Lake Marmara littoral core due to low unsupported 210Pb activities and high counting errors. Results indicated that Lake Marmara was dominated by benthic species throughout the sediment record, while Lakes Beyşehir and Uluabat shifted from a littoral-dominated system to one with increased pelagic species abundance. In all cores there was a stronger response to longer-term (decadal) and pronounced water-level changes than to short-term (annual-biennial) and subtle changes. It was also noted that degree of alteration in proxies differed between lakes, through time and among pelagic-littoral areas, likely emphasising differences in depositional environments and/or resolution of sampling and effects of other stressors such as eutrophication. Our results highlight lake-specific changes associated with water-level fluctuations, difficulties of conducting studies at required resolution in lakes with rather mixed sediment records and complexity of palaeolimnological studies covering recent periods where multiple drivers are in force. They further emphasise the need to include instrumental records when interpreting effects of recent water-level changes from sediment core data in large shallow lakes.