How does sea ice influence δ 18 O of Arctic precipitation ?

This study investigates how variations in Arctic sea ice and sea surface conditions influence δ18O of presentday Arctic precipitation. This is done using the model isoCAM3, an isotope-equipped version of the National Center for Atmospheric Research Community Atmosphere Model version 3. Four sensitivity experiments and one control simulation are performed with prescribed sea surface temperature (SST) and sea ice. Each of the four experiments simulates the atmospheric and isotopic response to Arctic oceanic conditions for selected years after the beginning of the satellite era in 1979. Changes in sea ice extent and SSTs have different impacts in Greenland and the rest of the Arctic. The simulated changes in central Arctic sea ice do not influence δ18O of Greenland precipitation, only anomalies of Baffin Bay sea ice. However, this does not exclude the fact that simulations based on other sea ice and sea surface temperature distributions might yield changes in the δ18O of precipitation in Greenland. For the Arctic, δ18O of precipitation and water vapour is sensitive to local changes in sea ice and sea surface temperature and the changes in water vapour are surface based. Reduced sea ice extent yields more enriched isotope values, whereas increased sea ice extent yields more depleted isotope values. The distribution of the sea ice and sea surface conditions is found to be essential for the spatial distribution of the simulated changes in δ18O.


Introduction
Records of stable water isotopologues from polar ice cores have been widely used to reconstruct past climate variability. Since the pioneering work by Dansgaard (1964), the understanding of stable water isotopologues as a proxy for temperature has significantly advanced. It has become clear that the isotopic composition of precipitation is a complex signal, influenced by both local and regional 20 climate conditions (Vinther et al., 2010;Steen-Larsen et al., 2011;Sjolte et al., 2011;Sodemann et al., 2008b;White et al., 1997;Johnsen et al., 1989). The isotopic composition of the precipitation is an integrated signal of the conditions along the moisture transport pathway from source to deposition.
1 As a result, there is a need for a detailed process-based understanding of the factors that can alter the isotopic composition of the transported moisture. 25 Studies using models, ice cores, snow and water vapour measurements have investigated the physical and dynamical processes influencing the isotopic composition of precipitation. Variations in local Greenland temperatures, conditions at source regions and atmospheric circulation all influence the isotopic composition of Greenland precipitation (Steen-Larsen et al., 2011;Bonne et al., 2014;Sodemann et al., 2008a, b;Sjolte et al., 2011;Vinther et al., 2010). 30 Several model studies highlight sea ice changes as important for understanding changes in the isotopic composition of precipitation. Sea ice changes in the Arctic were investigated during Dansgaard-Oeschger events (Li et al., 2010) and for exceptionally warm climates (Sime et al., 2013). For Antarctica, the impact of sea ice changes were studied using idealized reductions of the circular shaped sea ice cover (Noone, 2004). None of these model studies investigate sea ice perturbations comparable 35 to present-day observations. Measurements from ice cores spanning this period suggest that sea ice changes can influence the isotopic composition of precipitation (Divine et al., 2011;Opel et al., 2013;Ku et al., 2012;Fauria et al., 2010).
A study of idealised changes of Antarctic sea ice show a non-uniform spatial distribution of the modelled isotopic response over Antarctica (Noone, 2004). The heterogeneity of the response is 40 suggested to reflect the existence of different processes driving local and long range moisture transport to coastal and high elevation regions of Antarctica. Due to differences in the configuration of landmasses, open ocean and sea ice, it is difficult to directly transfer findings of Noone (2004) from Antarctic to the Arctic.
The impact of changes in sea ice and connected sea surface temperatures (SST) of the Arctic 45 ocean were studied by Sime et al. (2013). The sea ice conditions were created using an experiment where a coupled climate model was forced by respectively 2×, 4× and 8 × CO 2 . Hereafter the sea ice and SST conditions were used to force the applied atmospheric isotope models. Differences in the configurations of sea ice extent and SST were found to be essential for the resulting large variability in the isotope-temperature slope of 0.1 − 0.7 ‰/C for the Greenland ice sheet. While 50 these CO 2 changes used by Sime et al. (2013) do not allow direct comparison with present-day Arctic conditions, the results highlight processes that might be important for present day climate.
The recent decades of rapid Arctic sea ice decline provides an interesting opportunity to study how δ 18 O responds to realistic changes of sea ice and sea surface temperatures of present-day climate.
We here present results from isoCAM3 model simulations forced with observed Arctic sea ice and 55 sea surface temperature (SST) conditions derived from observations. This paper will address how the sea ice and sea surface conditions influence the δ 18 O in precipitation in the Arctic, and the role of the spatial configuration of the sea surface changes. The structure of the paper is as follows; (1) The model and experiments are described, (2) Results of the simulations are presented, (3) The influence of atmospheric moisture processes is discussed. The simulations of the isotopic composition of precipitation and water vapour in this study are conducted with isoCAM3. This is an atmospheric general circulation model (AGCM) enabled with the ability to trace the various species of water isotopologues. The model is based on the Community 65 Atmosphere Model version 3 (CAM3) (Collins et al., 2006), and the isotope module was developed by David Noone, University of Colorado. More details of isoCAM3 can be found in Noone and Sturm (2010) The model isoCAM3 has been applied in several studies that investigated the isotopic response to past climate changes (Tharammal et al., 2013;Speelman et al., 2010;Sturm et al., 2010;Pausata et al., 2011;Liu et al., 2012;Sewall and Fricke, 2013;Liu et al., 2014).

70
The horizontal resolution of the model is T85 (∼ 1. The solar constant is set to 1365 (W m −2 ) and orbital configurations are set to the year 1850.

Ensemble design
We perform a set of four sensitivity experiments and one control simulation to investigate how ob- We force the model isoCAM3 with an annual cycle of monthly mean SST and sea ice conditions 90 obtained from ERA-Interim (Dee et al., 2011). This annual cycle goes from April to March thus spanning the full sea ice cycle related to the selected cases of September sea ice extent. Hereafter the model runs for 15 years (following one year of spin up) with repeated annual cycle. All re-analysis data are interpolated bilinearly from the ERA-Interim (1 • x 1 • ) to the CAM3 T85 resolution, and hereafter checked for consistency.    Screen et al. (2013b) for further discussion on this topic).
Therefore, we chose that these experiments are based on both changes in sea ice and SST. A masking of the SST data is applied to eliminate remote influences from extra-polar climate patterns (e.g.

Atmospheric moisture processes
The δ 18 O p response to sea ice changes (Fig. 4 ) shows that the response is predominantly local, yet  While the anomalies of vapour and precipitation at the same location do not have to be linked, it still suggest that the anomalies of precipitation are not connected to changes in air masses and large scale transport but rather to local changes. This hypothesis is supported by observational studies of the impact of Arctic sea ice changes on the isotopic composition of moisture (Klein et al., 2015;Kopec et al., 2016). Changes in evaporation of 230 local ocean water have also been suggested by modelling studies as important for sea ice induced changes in δ 18 O p (Sime et al., 2013;Noone, 2004). Furthermore, an analysis of future warming in the Arctic using state-of-the-art climate models showed changes in the hydrological cycle due to Arctic warming and sea ice changes (Bintanja and Selten, 2014). In that study it was found that moisture inflow from lower latitudes played a minor role, and the changes in the hydrological cycle 235 were mainly caused by strongly intensified local surface evaporation.
An alternative explanation for the simulated changes in δ 18 O v and δ 18 O p is that the changes occur as result of changes in air mass characteristics. Reductions in the poleward temperature gradient would reduce the cooling and condensation that air masses experience during the northward transport. This would cause isotopic composition of the air masses to be less depleted. In this study, the 240 sea surface conditions effect on Arctic warming is clearly seen on the simulated surface air temperature (T 2m ) (see Fig. 3). Additionally also the vertical cross sections ( fig. 8 and 9)

Influence on Greenland precipitation
Changes in the isotopic composition of Greenland precipitation are of special interest due to the ice core research sites in this region. Interestingly, none of the sea ice perturbation experiments in this American heat wave, transitions in the Arctic Oscillation and transport of warm air and vapour via an atmospheric river (Neff et al., 2014;Bonne et al., 2014). Forcing the model with only oceanic conditions can thus not create a similar atmospheric-induced warming.
In contrast to the results of this study, Sime et al. (2013) simulated 2 − 3‰ changes in central Greenland δ 18 O p for extremely warm climates with SST and sea ice conditions created from coupled 285 model experiment forced by large increases in CO 2 . The main differences between the simulations in this study and in the study by Sime et al. (2013) are related to the distribution and magnitude of sea ice and SST changes especially near northern Greenland.
In the study by Sime et al. (2013) sea ice and SST changes also occur in the region north of Greenland. Also the magnitude of Arctic SST anomalies are 8−10 • C whereas the simulations in this 290 study have anomalies of 3 − 5 • C. These differences are compelling as our experiment "2012" with the largest prescribed SST anomalies and sea ice changes also is the only experiment that simulates a regional isotopic response. This indicates that the magnitude of SST changes might control not only the amount of local evaporation, but also the regional extent of the isotopic response. Hence, it is possible that the simulated changes of δ 18 O p by (Sime et al., 2013) have a regional extent due to 295 the same reasons as experiment "2012".
Warming of the lower troposphere and associated weakening of the inversion layer might be important in controlling the extent of the isotopic response. As sea ice removal is connected to intense warming of the lower troposphere (Screen et al., 2012;Deser et al., 2010), it could be speculated that this warming is controlling the extent of the isotopic response. This would be possible as a weaker 300 inversion layer allows atmospheric convection, and Abbot and Tziperman (2008) have shown that this can occur at high-latitudes in sea ice free regions in winter. Further investigation of the mechanism causing this change requires further idealized experiments following a similar to design to Noone (2004), so that a systematic investigation of the atmospheric processes influencing the isotopic composition of moisture is possible.

Conclusions
The aim of this study was to investigate whether changes in sea ice and sea surface temperatures derived from observed anomalies can influence the isotopic composition of precipitation in the Arctic.
Results are presented from isoCAM3 an isotope-equipped AGCM, forced with different distributions of Arctic sea ice changes and associated SST from the ERA-interim re-analysis product. These 310 simulations show that changes in sea ice and sea surface conditions influence the isotopic composition of Arctic precipitation with regional changes of δ 18 O p of up to 3‰ in the Barents Sea region.
However, no changes are found for Greenland; a region relevant for isotope records from ice cores.
For all experiments it is found that regions of increased ( Previous studies have shown that large changes in the state of sea ice and SST conditions influences the isotope composition over Greenland (Sime et al., 2013) and Antarctica (Noone, 2004) but 330 this study is the first model experiment to show that minor (relative to Sime et al. (2013)) perturba-tions in the sea ice cover and SST under present-day climate conditions can yield significant changes in the isotopic composition of precipitation in the Arctic, while at the same time not changing conditions in Greenland.
Acknowledgements. We thank the two anonymous reviewers for helpful comments and suggestions.The re- Annual mean δ 18 O p for the CTRL run is compared to observations of present day annual mean δ 18 O from Greenland ice cores (Vinther et al., 2010). Fig. 10 show that the isoCAM3 model has an annual mean positive bias of δ 18 O. No change in annual mean precipitation is found for each of the experiments compared to the CTRL run as shown in Fig. 11. 345 Figure 11. Annual mean anomalies of precipitation Anomalies for the four simulations compared to the CTRL run.

21
The spatial distributions of the anomalies of δ 18 O v at the 950 hPa level and 700 hPa level ( Fig.   12 and 13) show that the anomalies of δ 18 O v are mostly found at surface levels for the entire Arctic region. 22