Search for: All records

Accurate measurement of net radiation in the high-latitude Arctic regions is challenging since rain and snow events often introduce substantial measurement errors. To reduce the precipitation-induced measurement errors of downward radiation, customized data-driven methods are developed to reconstruct downward radiative fluxes from the biased radiation measurements. This study uses four years of field data across ten plots covered with forest, trees, and tundra in the Polar Urals from July 2018 to July 2022. Rain and snow on the radiometers absorb and block shortwave radiation and emit longwave radiation, leading to underestimation of downward shortwave and overestimation of downward longwave radiation. Snow causes more errors than rain. Seasonal variation of reconstructed net radiation for three dominant vegetation types indicates that their differences are most pronounced in April and least in September. Furthermore, forest and tree plots consistently exhibit higher magnitudes of net radiation and longer seasons of positive net radiation than tundra plots. This study advances methodologies for reconstructing corrupted net radiation data in the Arctic and offers insights into the variability of net radiation patterns within the forest-tundra ecotone. 

Abstract Applications of process‐based models (PBM) for predictions are confounded by multiple uncertainties and computational burdens, resulting in appreciable errors. A novel modeling framework combining a high‐fidelity PBM with surrogate and machine learning (ML) models is developed to tackle these challenges and applied for streamflow prediction. A surrogate model permits high computational efficiency of a PBM solution at a minimum loss of its accuracy. A novel probabilistic ML model partitions the PBM‐surrogate prediction errors into reducible and irreducible types, quantifying their distributions that arise due to both explicitly perceived uncertainties (such as parametric) or those that are entirely hidden to the modeler (not included or unexpected). Using this approach, we demonstrate a substantial improvement of streamflow predictive accuracy for a case study urbanized watershed. Such a framework provides an efficient solution combining the strengths of high‐fidelity and physics‐agnostic models for a wide range of prediction problems in geosciences. 

Abstract This study develops a novel general framework to project the permafrost fate with rigorous uncertainty quantification to assess dominant sources. Borehole temperature records from three sites in the Russian western Arctic are used to constrain the uncertainty of a high‐fidelity freeze‐thaw model. Projections from 9 Global Climate Models (GCM) are stochastically downscaled to generate future trajectories of surface ground heat flux. Under the two emission scenarios SSP2‐4.5 and SSP5‐8.5, the projected average thawing depths by 2100 vary from 0.4 to 14.4 m or 2.1 to 17.7 m, and the increase in the top 10 m average temperature from 2015 to 2100 is 1.2–2.7°C or 1.9–3.0°C. The results show that the freeze‐thaw model uncertainty can sometimes dominate over that of GCM outputs, calling for site‐specific information to improve model accuracy. The framework is applicable for understanding permafrost degradation and related uncertainties at larger scales.