Many paleoclimate reconstructions rely on geochemical proxies to determine past environmental changes. However, in Mediterranean Sea surface sediments, certain sea surface temperature (SST) proxies often yield values that are offset from the expected SST (

Many paleoclimate reconstructions rely on geochemical proxies to determine past environmental changes. However, in Mediterranean Sea surface sediments, certain sea surface temperature (SST) proxies often yield values that are offset from the expected SST (

Many water-stressed regions of the globe have a highly seasonal precipitation regime. However, seasonality in the past and under changing climates is little studied. Submonthly records of sclerochronological δ18O and δ13C values of Melanopsis shells from the Jordan River Dureijat archaeological site (JRD) in the upper Jordan River valley presented here document the hydrology of paleo-Lake Hula. These records were assessed for changes in seasonal hydrology in the lake and compared with modern shells collected from present-day waterbodies in northern Israel and with models of δ18Oshell. Results from shells in sediments dating from the last glacial maximum (LGM) to the Bølling-Allerød imply changes in waterbody size that qualitatively parallel changes in the late Pleistocene Lake Lisan levels; Hula Lake was well buffered when Lake Lisan stood at a high stand and poorly buffered when water levels were lower. Furthermore, data from shells dated to the LGM suggest inflowing water with lower δ18O values than local rainfall, providing evidence for a greater proportion of snow in the catchment than today. Reconstruction of water δ18O and mixing-model calculations suggest that snowmelt contribution to spring water during the LGM may have been more than twice the amount in the modern-day catchment.

<p>Dust deposition can trigger phytoplankton growth in oligotrophic low nutrient low chlorophyl (LNLC) regions by providing essential nutrients to the surface ocean. As LNLC regions comprise 60% of the global ocean, dust fertilisation and potential subsequent increased downward carbon export could affect the strength of the biological carbon pump considerably. Additionally, ballasting effects of large dust particles could enhance downward carbon export even further, independent from fertilisation effects. However, compared to high nutrient low chlorophyl (HNLC) regions, the biogeochemical effect of dust deposition and its sensitivity to future climate change is less well understood for LNLC regions. For the LNLC Mediterranean Sea mesocosm experiments and satellite data suggest that some, but not all, dust events lead to increased primary production. However, the exact relationship between dust deposition, productivity and carbon export remains unresolved.  </p> <p>Here, we aim to identify and quantify the relationships between Saharan dust deposition (deposition mode, dust source), phytoplankton response (changes in community composition, phytoplankton vs heterotrophic bacterial growth) and carbon export in the eastern Mediterranean Sea by studying an exceptional high-resolution, 30-year sediment-trap time series of settling Saharan dust particles and phytoplankton remains (partly at 500m, 1500m, and 2500m water depth), combining sedimentological, biogeochemical, and remote sensing techniques. We here present a combined record of dust and organic matter fluxes for one full year of the time series (April 2017 to May 2018, 2200m water depth). Furthermore, the response of specific phytoplankton groups to dust input as well as the input of terrestrial plant material associated with desert dust is determined based on the presence and distribution of lipid biomarkers in the trap material.</p> <p>Dust fluxes vary substantially over this one-year period, but peaks occur during spring 2017 and 2018, summer 2017, as well as some smaller, less pronounced peaks during autumn 2017. Some of these dust events indeed correspond to increased fluxes of lipid biomarkers, suggesting a relationship between dust input and enhanced sinking of organic material. However, due to the depth of trap deployment, the record does not allow to differentiate between the influence of dust input as fertiliser or as ballasting effect. This will later be assessed by comparing biomarker records from sediment traps from different depths representing the surface and deep ocean. Nevertheless, the lipid biomarkers representing different phytoplankton groups (e.g., long-chain alkenones for coccolithophores, 23,24-dimethylcholesta-5,22E-dien-β-ol for diatoms, dinosterol for dinoflagellates, long-chain diols for eustigmatophytes) do not show a uniform response to dust input, indicating that the response of these phytoplankton groups depends on different conditions. Moreover, some dust events do not seem to trigger any phytoplankton response at all as they do not coincide with enhanced biomarker fluxes. This indicates that other factors such as dust source, deposition mode and/or trophic state of the surface ocean determine whether dust input triggers enhanced export of organic material or not. Differences in grain-size distribution and terrestrial plant content (indicated by terrestrial plant biomarkers) indeed suggest that the observed contrasting response might be due to differences in dust source and composition. </p>

<p>Ocean currents can transport sinking particles hundreds of kilometers from their origin at the ocean surface to their burial location, resulting in an offset between sea surface temperatures (SSTs) above the burial site and the particle’s origin. Quantifying this offset in particles carrying molecules used in SST proxies can reduce uncertainty in paleoclimate reconstructions. In the Mediterranean Sea, where δ<sup>18</sup>O<sub>foraminifera</sub>, U<sup>K’</sup><sub>37</sub>- and TEX<sub>86</sub>-based SSTs can exhibit large offsets from surface conditions, understanding the possible contribution of lateral transport to proxy bias can provide additional insight when interpreting paleoclimate records.</p><p>In this study, Lagrangian particle tracking experiments are performed using the NEMO flow field to simulate transport and allow for a quantitative estimate of transport bias. The model determines the ocean surface origin locations of foraminifera and sedimentary particles that carry alkenones or GDGTs to compare with surface sediment datasets for δ<sup>18</sup>O<sub>foraminifera</sub>, U<sup>K’</sup><sub>37</sub> and TEX<sub>86</sub>, respectively. A range of sinking speeds appropriate for the export of organic matter (6, 12, 25, 50, 100, 250, and 500 m/d) is used in the model to represent different export modes (i.e., individual coccoliths, coccospheres, aggregates), where the three fastest sinking speeds can also represent sinking foraminifera. Results show that lateral transport bias is generally small within the Mediterranean Sea and cannot explain the large offsets in proxy-based SST reconstructions in this basin.</p>