Simulating asymmetric growth and retreat of Hardangerjøkulen ice cap in southern Norway since the mid-Holocene

Abstract

Changes to the volume of glaciers and ice caps currently amount to half of the total cryospheric contribution to sea-level rise and are projected to remain substantial throughout the 21st century. To simulate glacier behavior on centennial and longer time scales, models rely on simplified dynamics and tunable parameters for processes not well understood. Model calibration is often done using present-day observations, even though the relationship between parameters and parametrized processes may be altered for significantly different glacier states. In this study, we simulate the evolution of the Hardangerjøkulen ice cap in southern Norway from the mid-Holocene through the Little Ice Age (LIA) to the present-day. For both the calibration and transient experiments, we run an ensemble using a two-dimensional ice flow model with local mesh refinement. For the Holocene, we apply a simple surface mass balance forcing based on climate reconstructions. For the LIA until 1962, we use geomorphological evidence and measured outlet glacier positions to find a mass balance history, while from 1963 until today we use direct mass balance measurements. Given a linear climate forcing, we find that Hardangerjøkulen grew from ice-free conditions in the mid-Holocene, to its maximum LIA extent in a highly non-linear fashion. During the fastest stage of growth (2200-1200 BP), the ice cap tripled its ice volume over only 1000 years. We also reveal an intriguing spatial asymmetry during advance and retreat; the western ice cap and the northern outlet glacier Midtdalsbreen grow first and disappear first. In contrast, the eastern part, including the northeastern outlet glacier Blåisen, grows last and disappears last. Furthermore, volume and area of several outlet glaciers, as well as of the entire ice cap, vary out-of-phase for multiple centuries during the late Holocene, before varying in-phase approaching the LIA. We relate this to bed topography and the mass balance-altitude feedback, and challenge canonical linear assumptions between ice cap extent and glacier proxy records. Thus, we provide new insight into long-term dynamical response of ice caps to climate change, relevant for paleoglaciological studies and future predictions. Our model simulates ice cap extent and outlet glacier length changes from the LIA until today that are close to observations. We show that present-day Hardangerjøkulen is extremely sensitive to surface mass balance changes, mainly due to a strong mass balance-altitude feedback for the gently sloping surface topography of the ice cap.