Firn aquifers play an important role in the mass balance of the Greenland Ice Sheet, but changes in their water storage are challenging to measure. Here, we explore the feasibility of passive radar sounding to monitor changes in the firn aquifer water table level throughout the year. Our simulations consider the time-varying environmental challenges for passively monitoring fluctuations in firn aquifer water table height. Our results suggest that passive sounding could be a novel observational tool for quasi-continuous monitoring of firn aquifer properties in geographic regions of interest.

Ice-penetrating radar (IPR) instruments are a widely used tool to understand the structure and dynamics of Earth’s ice sheets and glaciers. Originally primarily designed to image the bedrock beneath ice, IPR systems are now being used for a wider range of scientific investigations. At the same time, new hardware architectures and customized radar systems are emerging. This combination of factors makes it worth re-visiting common assumptions about the noise characteristics of IPR systems and the phase coherence of measured reflections. In this work, we explore what it means for an IPR instrument to be coherent, how architectural choices in the design of software-defined radio-based instruments may impact phase coherence, and what instrument specifications should be considered for IPR applications that rely on measurements of small phase changes.

Meltwater percolation and refreezing in porous firn is difficult to observe, but plays a key role in controlling the rate of melt-water runoff from the Greenland Ice Sheet. Coherent change detection (CCD) with repeated airborne ice-penetrating radar acquisitions is a promising method for resolving these processes. However, existing surveys were not designed for interferometric processing, and CCD has not yet been applied to near-surface ice-penetrating radar data. Here, we develop a workflow for estimating the coherence between repeat acquisitions by the Operation IceBridge Accumulation Radar. We demonstrate that stratigraphic structures such as ice slabs and ice layers maintain good coherence even over a temporal baseline of one year. However, porous firn with heterogeneous ice lensing has poor coherence, suggesting that tighter baseline control and improved co-registration and motion compensation methods would be required to reliably observe subsurface change in the percolation zone.

Orbital radar sounding systems could significantly improve our understanding of dynamic processes at ice sheet margins by enhancing the spatial and temporal coverage of subglacial observations. However, surface crevasses in these fast-flowing regions may contribute significant radar clutter that can impede detection of the ice sheet bed. It is currently unclear if particular radar center frequency choices might enhance or mitigate this clutter. To investigate this question, we analyze crevasse clutter returns in airborne radar sounding data collected at 195 MHz, 300 MHz, and 750 MHz. Additionally, we rescale the derived empirical angular scattering functions to simulate crevasse clutter in orbital geometries. In this data set, we find that, unlike rough surface clutter, crevasse clutter is reduced at higher frequencies. However, the small angles of incidence at orbital altitudes lead to sufficient clutter at all frequencies to dominate over the theoretical bed echo power. Our results suggest that surface crevassing may not necessarily limit the viability of UHF frequencies for orbital sounding. However, synthetic aperture focusing and narrow cross-track antenna beam patterns will be crucial regardless of system center frequency.

Perennial firn aquifers may play an important role in modulating the surface mass balance and ice dynamics of the Greenland Ice Sheet by storing surface meltwater in a porous firn layer. Ice-penetrating radar has been shown to be an effective tool for mapping the spatial extent and thickness of these subsurface water bodies. However, quantitative estimates of aquifer properties from radar observations require extensive a priori knowledge of the local firn structure and electrical properties of the aquifer water. Here, we develop a Markov Chain Monte Carlo joint inversion of radar reflectivity and attenuation to simultaneously estimate firn structure and aquifer properties and their associated uncertainty. We show that using this inversion method, the only additional field measurement required to robustly and accurately constrain the total water storage is the conductivity of the aquifer water. However, the inversion is largely insensitive to firn structure parameters, and aquifer properties like thickness and bulk porosity are poorly constrained on an individual basis due to the large number of solutions that map to the same total water storage. Regardless, our results suggest that, when combined with field measurements of water conductivity, ice-penetrating radar surveys can be an effective method for remotely characterizing kilometer-scale spatial and temporal variations in firn aquifer storage capacity.

Surface crevasses impact ice sheet mass loss by initiating hydrofracturing and calving at the margins and transporting supraglacial meltwater to the subglacial drainage sys-tem. This process subsequently modulates basal sliding and glacier motion. However, the development of robust models for calving and hydrofracture has been limited by a lack of field observations of crevasse formation and geometry. In this paper, we analyze a two-year Multi-Input Multi-Output Autonomous Phase-Sensitive Radio-Echo Sounder (MIMO ApRES) dataset collected at Store Glacier in West Greenland, which documents the formation of a crevasse that opened under the instrument. We present methods for processing the data as well as identifying and removing artifacts, including clipping, radio frequency interference (RFI), receiver failure events such as elevated thermal noise, and signal leakage between channels. Specifically, we perform a mean squared error (MSE) analysis, clipping detection and quantification, and calculations of total power over time in the frequency domain and the time domain. After characterizing and min-imizing these artifacts, we find that the bottom of a crevasse can be detected in the processed images. Our results suggest that, with appropriate data processing, the MIMO ApRES is a promising geophysical system for investigating future crevasse evolution.

Englacial layering is a ubiquitous feature of ice penetrating radar images of continental ice sheets. These reflections occur at subsurface interfaces of changing density, conductivity, or crystal orientation fabric and can record the past atmospheric conditions and flow history of a region. As a result, both the physical cause of this layering and its geometry are of scientific interest. We develop a method for simulating englacial radio stratigraphy using measurements of density and conductivity from deep ice cores and show that it can reproduce coincident radar measurements. We then extend this method to two dimensions to simulate physically realistic synthetic radargrams. Our model incorporates both geophysical parameters and radar system parameters including center frequency, bandwidth, impulse response, aircraft attitude, and post-processing. As a result, our model can be used to generate labeled training data for machine learning layer tracking algorithms, connect layers observed by different radar systems to ice core stratigraphy, or as a forward model for testing hypotheses for the formation and evolution of the geometric and chemical properties of englacial layers away from deep drilling sites.

Quantifying uncertainty in subglacial topography measured with airborne radio-echo sounding (RES) is important for studying ice sheet dynamics and constraining future ice sheet behavior. Measurement accuracy is influenced by bed geometry. However, measurement uncertainties from off-nadir reflections are insufficiently characterized for quantifying topographic measurement uncertainty from off-nadir reflections. In this paper, we use a clutter simulator to quantify this source of uncertainty. We find that measurement uncertainty from clutter is spatially correlated and is strongly dependent on topographic roughness. This method could be used to estimate measurement uncertainty in RES measurements and enable more robust uncertainty quantification in bed topography.

Orbital radar sounding of terrestrial ice sheets is an area of increasing research interest with mission concepts at 45 MHz, P-Band, and L-Band under development. However, large uncertainties remain in impact of glacial conditions and platform altitude on their link budgets. Here, we present a collection of empirically and glaciologically informed constraints on orbital sounder link budgets using airborne radar sounding data. We also analyze the effects of geometric spreading and englacial water. Finally, we discuss link-budget considerations for investigations beyond bed mapping including observing basal reflectivity, englacial hydrology, ice-shelf thickness, englacial layers, and estimating vertical ice velocity.

The formation of impermeable ice layers within Greenland's firn can significantly increase surface meltwater runoff by capping percolation, thus increasing the continent's immediate contributions to sea level rise. Detection of these layers by airborne radar sounding would permit large scale assessment of their spatial coverage and temporal evolution. We present an electromagnetic forward model for radar scattering in dry firn, as well as a statistical model of firn density profiles, which together allow us to robustly simulate airborne radar sounding measurements in the ice sheet near-surface. We use these models to simulate the response of the University of Kansas Accumulation Radar to refrozen ice layers thinner than the radar vertical resolution. Our results suggest that continuous ice layers are detectable as anomalously bright reflections within the firn, so long as the background density does not exceed 0.72 g/cm3. We find that approximately 81% of single ice layers thicker than 2 cm are detectable, as well as over 90% of all multi-layer configurations. This suggests that the Accumulation Radar is an effective tool for studying the spatial and temporal coverage of thin ice layers in Greenland's firn, but such a survey would likely still underestimate the total areal extent.

Radar sounding is a powerful tool for constraining subglacial conditions, which influence the mass balance of polar ice sheets and their contributions to global sea-level rise. A satellite-based radar sounder, such as those successfully demonstrated at Mars, would offer unprecedented spatial and temporal coverage of the subsurface. However, airborne sounding studies suggest that poorly constrained radar scattering in polar firn may produce performance-limiting clutter for terrestrial orbital sounders. We develop glaciologically constrained electromagnetic models of radar interactions in firn, test them against in situ data and multifrequency airborne radar observations, and apply the only model we find to be consistent with observation to assess the implications of firn clutter for orbital sounder system design. Our results show that in the very high-frequency (VHF) and ultrahigh-frequency (UHF) bands, radar interactions in the firn are dominated by quasi-specular reflections at the interfaces between layers of different densities and that off-nadir backscatter is likely the result of small-scale roughness in the subsurface density profiles. As a result, high frequency (HF) or low VHF center frequencies offer a significant advantage in near-surface clutter suppression compared to the UHF band. However, the noise power is the dominant constraint in all bands, so the near-surface clutter primarily constrains the extent to which the transmit power, pulselength, or antenna gain can be engineered to improve the signal-to-noise ratio. Our analysis suggests that the deep interior of terrestrial ice sheets is a difficult target for orbital sounding, which may require optimizations in azimuth processing and cross-track clutter suppression which complement existing requirements for sounding at the margins.

Radar sounding of ice from orbit has been successful on Mars [1], is planned for the Galilean satellites [2], and is attractive for earth [3] as a complement to current airborne instruments [4], but of major concern is the poorly constrained but potentially seriously limiting contribution of firn clutter [5]. To inform this issue, we analytically model electromagnetic scattering in the upper 100 meters of the ice column for continental ice sheets and evaluate the effects of variable platform altitude, frequency, and range resolution on clutter power. Our results show that volume scattering from air inclusions is insignificant and unlikely to constrain deep ice sounding. Rather, firn scattering is dominated by quasispecular reflections from layers of varying density which, at orbital altitudes, may contribute significantly to clutter due to the small angles of illumination. This layer clutter can be mitigated by a careful choice of range resolution for center frequencies below 200 MHz, but is practically unavoidable above 250 MHz. Firn layer clutter is likely to significantly constrain UHF orbital ice sounding, making a VHF instrument the more practical choice.