• Impacts of storm on sea ice: from case study to climate scale analysis

      Peng, Liran; Zhang, Xiangdong; Collins, Richard; Fochesatto, Javier; Polyakov, Igor (2019-12)
      Recent studies have shown that intense and long-lasting storms potentially facilitate sea ice melting. Under the background of extratropical storm tracks poleward shift, significant reductions of Arctic sea ice coverage, and thinning of sea ice thickness over the last several decades, a better understanding on how storms impact sea ice mass balance is obviously of great importance to better predict future sea ice and the Arctic climate changes. This thesis presents a multi-scale study on how storms impact sea ice, consisting of three different parts of the effort. In the first part, we examined the impacts of the 2016 summer intense storm on sea ice changes over the Chukchi Sea using ship-borne observations. The results show that the intense storm can accelerate ice melt through enhanced upper-ocean mixing and upward heat transport. The satellite-observed long-term sea ice variations potentially can be impacted by many factors. In the second part, we first explore key physical processes controlling sea ice changes under no-storm condition. We examined and compared results from 25 sensitivity experiments using the NCAR's Community Earth System Model (CESM). We found that sea ice volume, velocity, and thickness are highly sensitive to perturbed air-ice momentum flux and sea ice strength. Increased sea ice strength or decreased air-ice momentum flux causes counter-clockwise rotation of the transpolar drift, resulting in an increase in sea ice export through Fram Strait and therefore reduction of the pan-Arctic sea ice thickness. Following four tracers released over the Arctic, we found the sea ice thickness distributions following those tracers are broader over the western Arctic and becomes narrower over the eastern Arctic. Additionally, thermodynamic processes are more dominant controlling sea ice thickness variations, especially over periphery seas. Over the eastern Arctic, dynamic processes play a more important role in controlling sea ice thickness variation. Previous studies show that thin ice responds to external perturbations much faster than the thick ice. Therefore, the impacts of storms on sea ice are expected to be different compared with the western/eastern Arctic and the entral/periphery seas. In the third part, we conduct a new composite analysis to investigate the storm impact on sea ice over seven regions for all storms spanning from 1979 to 2018. We focused on sea ice and storm changes over seven regions and found storms tend to have different short-term (two days before and after storm passage), mid-term (one-two weeks after storm passage), and long-term (from 1979 to 2018) impact on sea ice area over those regions. Over periphery seas (Chukchi, East Siberian, Laptev, Kara, and Barents Seas), storms lead to a short-term sea ice area decrease below the climatology, and a mid-term sea ice increase above the climatology. This behavior causes sea ice area to have a small correlation with the storm counts from 1979 to 2018, which suggest that storms have a limited long-term impact on sea ice area over periphery seas. Both the short term and mid-term storm impacts on sea ice area are confined within a 400 km radius circle with maximum impacts shown within a 200 km radius circle. Storms over the western Arctic (Chukchi, East Siberian, and Laptev Seas) have a stronger short-term and mid-term impact on sea ice area compared with the Eastern Arctic (Barents and Kara Seas). Storms over both Atlantic and Pacific entrance regions have a small impact on sea ice area, and storms over the Norwegian, Iceland, and Greenland Seas have the smallest impact on the sea ice area. Compared to the periphery seas, storms tend to have a stronger long-term impact on sea ice area over the central Arctic. The correlation coefficients between the storm count and sea ice area exceed 0.75.
    • Modelling investigation of interaction between Arctic sea ice and storms: insights from case studies and climatological hindcast simulations

      Semenov, Alexander; Zhang, Xiangdong; Bhatt, Uma; Hutchings, Jennifer; Mölders, Nicole (2019-05)
      The goal of this study is to improve understanding of atmosphere, sea ice, and ocean interactions in the context of Arctic storm activities. The reduction of Arctic sea ice extent, increase in ocean water temperatures, and changes of atmospheric circulation have been manifested in the Arctic Ocean along with the large surface air temperature increase during recent decades. All of these changes may change the way in which atmosphere, sea ice, and ocean interact, which may in turn feedback to Arctic surface air warming. To achieve the goal, we employed an integrative approach including analysis of modeling simulation results and conducting specifically designed model sensitivity experiments. The novelty of this study is linking synoptic scale storms to large-scale changes in sea ice and atmospheric circulation. The models were used in this study range from the regional fully coupled Arctic climate model HIRHAM-NAOSIM to the ocean-sea ice component model of the Community Earth System Model CESM and the Weather Research and Forecasting (WRF) model. Analysis of HIRHAM-NAOSIM simulation outputs shows regionally dependent variability of storm count with a higher number of storms over the Atlantic side than over the Pacific side. High-resolution simulations also reproduce higher number of storms than lower resolution reanalysis dataset. This is because the high-resolution model may capture more shallow and small size storms. As an integrated consequence, the composite analysis shows that more numerous intense storms produce low-pressure systems centered over the Barents-Kara-Laptev seas and the Chukchi-East Siberian seas, leading to anomalous cyclonic circulation over the Atlantic Arctic Ocean and Pacific Arctic Ocean. Correspondingly, anomalous sea ice transport occurs, enhancing sea ice outflow out of the Barents-Kara-Laptev sea ice and weakening sea ice inflow into the Chukchi-Beaufort seas from the thick ice area north of the Canadian Archipelago. This change in sea ice transport causes a decrease in sea ice concentration and thickness in these two areas. However, energy budget analysis exhibits a decrease in downward net sea ice heat fluxes, reducing sea ice melt, when more numerous intense storms occur. This decrease could be attributed to increased cloudiness and destabilized atmospheric boundary layer associated with intense storms, which can result in a decrease in downward shortwave radiation and an increase in upward turbulent heat fluxes. The sea ice-ocean component CICE-POP of Community Earth System Model (CESM) was used to conduct sensitivity experiment to examine impacts of two selected storms on sea ice. CICE-POP is generally able to simulate the observed spatial distribution of the Arctic sea-ice concentration, thickness, and motion, and interannual variability of the Arctic sea ice area for the period 1979 to 2011. However, some biases still exit, including overestimated sea-ice drift speeds, particularly in the Transpolar Drift Stream, and overestimated sea-ice concentration in the Atlantic Arctic but slightly underestimated sea ice concentration in the Pacific Arctic. Analysis of CICE-POP sensitivity experiments suggests that dynamic forcing associated with the storms plays more important driving role in causing sea ice changes than thermodynamics does in the case of storm in March 2011, while both thermodynamic and dynamic forcings have comparable impacts on sea ice decrease in the case of the August 2012. In case of March 2011 storm, increased surface winds caused the reduction of sea ice area in the Barents and Kara Seas by forcing sea ice to move eastward. Sea ice reduction was primarily driven by mechanical processes rather than ice melting. On the contrary, the case study of August 2012 storm, that occurred during the Arctic summer, exemplified the case of equal contribution of mechanical sea ice redistribution of sea ice in the Chukchi - East Siberian - Beaufort seas and melt in sea ice reduction. To understand the impacts of the changed Arctic environment on storm dynamics, we carried out WRF model simulations for a selected Arctic storm that occurred in March 2011. Model output highlight the importance of both increased surface turbulent heat fluxes due to sea ice retreat and self-enhanced warm and moist air advection from the North Atlantic into the Arctic. These external forcing factor and internal dynamic process sustain and even strengthen atmospheric baroclinicity, supporting the storm to develop and intensify. Additional sensitivity experiments further suggest that latent heat release resulting from condensation/precipitation within the storm enhances baroclinicity aloft and, in turn, causes a re-intensification of the storm from its decaying phase.
    • A study on the morphology of magnetic storms

      Sugiura, Masahisa (Geophysical Institute at the University of Alaska, 1955-04-20)
      The morphology of magnetic storms that has been investigated by S. Chapman since 1918 was further extended with more material as regards both the number of storms and the number of observatories, Three hundred and forty-six magnetic storms having sudden commencements were selected for the years 1902 to 19^5- These 3^6 storms were classified by a new method based on the most notable characteristic of the stormtime variation observed in low and middle latitudes, namely, a worldwide diminution in the horizontal force; hence the maximum diminution in the horizontal force averaged over these latitudes was used as a measure to indicate the intensity of magnetic storms. The 3^6 storms here selected were classified into three intensity groups: (l.) weak, (2) active, and {3 ) great storms.. The numbers of storms classed in these categories are 136, 136 and kj, respectively. In the present thesis the investigation on the 136 weak magnetic storms is described. (in the previous study made by Chapman forty storms of moderate intensity were used.) The number of magnetic observatories used in, the present study was also widely extended from eleven (in the previous work) to twentyfive. Seven observatories in the southern hemisphere distributed between geomagnetic latitudes 12° and 48° were included in these twentyfive observatories„ The geomagnetic latitudes of the eighteen northern observatories range from 20° to 80°, Several improvements were also made in the treatment of the magnetic data. One of the improvements is that hour-to-hour differences derived from the hourly values of the three magnetic elements, the horizontal force, declination and the vertical, force, were used, instead of the ordinary hourly values as given in observatory reports. With the 136 weak magnetic storms the storm-time variations of the three elements for the four pre-storm hours and the first seventy-two hours from the storm commencement were determined for eight groups of observatories, whose mean geomagnetic latitudes are 28° S, 21°, 28°, k2°, 52°, 59°, 65° and 80°; the first group being in the southern, the rest in the northern hemisphere. Thus the average features of the storm-time variations at various latitudes were able to be studied more closely. In determining these storm-time changes the daily variations on quiet days uncorrected for the non-cyclic variation were removed from the original data in order to allow for this latter variation. Then the disturbance local-time inequality for the first, second and third storm days was examined for each magnetic element for each of the groups of observatories. The vectograms of these variations were also drawn. Besides confirming, on the whole, the views expressed in Chapman's discussions on the storm-time as well as the disturbance local-time inequality, the present results revealed more detailed features of these variations at various latitudes. The disturbance local-time inequality for each element for each group of observatories was further studied for shorter intervals of storm-time, that is, for four 6-hour intervals in each of the first and second days and for three 8-hour intervals in the third storm day. The results were harmonically analyzed to determine the diurnal (2^-hour) and the semi-diurnal (12-hour) components of these variations. The diurnal component was illustrated by harmonic, dials, by which means the decay of the amplitude of these variations and the change of their phases with storm-time were clearly demonstrated. It was found that the phases of the disturbance diurnal equality in declination and the vertical force have certain definite relations with that in the horizontal force at each latitude, and that if the results for declination and the vertical force are combined -with those for the horizontal force with some appropriate modifications in their amplitudes and phases, and if such averages are further combined among the groups of observatories in low and middle latitudes, the averaged harmonic dial so obtained is much more regular than those for individual elements or for smaller groups. The rates of growth and decay of the storm-time change and the disturbance local-time inequality were compared., The results indicate that these two variations vary at rates that are materially different in their course. Detailed descriptions and discussions on these results, the final objects of the present study and plans for its future extension, are given in the present thesis.