High Resolution Modeling of the Water Cycle to Refine GRACE Signal Analysis in the Gulf of Alaska Drainage

Abstract:

A comprehensive study of the Gulf of Alaska (GOA) drainage basin was carried out to improve understanding of the coastal freshwater discharge (FWD) and surface mass balance (SMB) of glaciers. Coastal FWD and SMB for all glacier surfaces were modeled using a suite of physically based, spatially distributed weather, energy-balance snow/ice melt, soil water balance, and runoff routing models at a high resolution (1 km horizontal grid; daily time step). A 35 year hind cast was performed, providing complete records of precipitation, runoff, snow water equivalent (SWE) depth, evapotranspiration, coastal FWD and glacier SMB. Meteorological forcing was provided by the North American Regional Reanalysis (NARR), Modern Era Retrospective Analysis for Research and Applications (MERRA), and NCEP Climate Forecast System Reanalysis (CFSR) datasets. A fourth dataset was created by bias-correcting the NARR data to recently-developed monthly weather grids based on PRISM climatologies (NARR-BC). Each weather dataset and model combination was individually calibrated using PRISM climatologies, streamflow, and glacier mass balance measurements from four locations in the study domain. Simulated mean annual FWD into the GOA ranged from 600 km³ yr^-1 using NARR to 850 km³ yr^-1 from NARR-BC. The CFSR-forced simulations with optimized model parameters produced a simulated regional water storage that compared favorably to data from the NASA/DLR Gravity Recovery and Climate Experiment (GRACE) high resolution mascon solutions (Figure). Glacier runoff, taken as the sum of rainfall, snow and ice melt occurring on glacier surfaces, ranged from 260 km³ yr^-1 from MERRA to 400 km³ yr^-1 from NARR-BC, approximately one half of the signal from both glaciers and surrounding terrain. The large contribution from non-glacier surfaces to the seasonal water balance is likely not being fully removed from GRACE solutions aimed at isolating the glacier signal alone. We will discuss methods to use our simulations to forward-model the hydrology of the Gulf of Alaska region and minimize uncertainty in the partitioning of the hydrological signal. This study provides significant insight into the linkages between hydrological modeling and gravimetric measurements in mountain environments.