ERA5-Land (ERA5L) is a reanalysis product derived by running the land component of ERA5 at increased resolution. This study evaluates ERA5L soil temperature in permafrost regions based on observations and published permafrost products. We find that ERA5L overestimates soil temperature in northern Canada and Alaska but underestimates it in mid–low latitudes, leading to an average bias of −0.08 ^∘C. The warm bias of ERA5L soil is stronger in winter than in other seasons. As calculated from its soil temperature, ERA5L overestimates active-layer thickness and underestimates near-surface (<1.89 m) permafrost area. This is thought to be due in part to the shallow soil column and coarse vertical discretization of the land surface model and to warmer simulated soil. The soil temperature bias in permafrost regions correlates well with the bias in air temperature and with maximum snow height. A review of the ERA5L snow parameterization and a simulation example both point to a low bias in ERA5L snow density as a possible cause for the warm bias in soil temperature. The apparent disagreement of station-based and areal evaluation techniques highlights challenges in our ability to test permafrost simulation models. While global reanalyses are important drivers for permafrost simulation, we conclude that ERA5L soil data are not well suited for informing permafrost research and decision making directly. To address this, future soil temperature products in reanalyses will require permafrost-specific alterations to their land surface models.

Data assimilation plays a dual role in advancing the “scientific” understanding and serving as an “engineering tool” for the Earth system sciences. Land data assimilation (LDA) has evolved into a distinct discipline within geophysics, facilitating the harmonization of theory and data and allowing land models and observations to complement and constrain each other. Over recent decades, substantial progress has been made in the theory, methodology, and application of LDA, necessitating a holistic and in‐depth exploration of its full spectrum. Here, we present a thorough review elucidating the theoretical and methodological developments in LDA and its distinctive features. This encompasses breakthroughs in addressing strong nonlinearities in land surface processes, exploring the potential of machine learning approaches in data assimilation, quantifying uncertainties arising from multiscale spatial correlation, and simultaneously estimating model states and parameters. LDA has proven successful in enhancing the understanding and prediction of various land surface processes (including soil moisture, snow, evapotranspiration, streamflow, groundwater, irrigation and land surface temperature), particularly within the realms of water and energy cycles. This review outlines the development of global, regional, and catchment‐scale LDA systems and software platforms, proposing grand challenges of generating land reanalysis and advancing coupled land‒atmosphere DA. We lastly highlight the opportunities to expand the applications of LDA from pure geophysical systems to coupled natural and human systems by ingesting a deluge of Earth observation and social sensing data. The paper synthesizes current LDA knowledge and provides a steppingstone for its future development, particularly in promoting dual driven theory‐data land processes studies.

The Tibetan Plateau observatory (Tibet-Obs) of plateau scale soil moisture and soil temperature was established 10 years ago and has been widely used to calibrate/validate satellite- and model-based soil moisture (SM) products for their applications to the Tibetan Plateau (TP). This paper reports on the status of the Tibet-Obs and presents a 10-year (2009–2019) surface SM dataset produced based on in situ measurements taken at a depth of 5 cm collected from the Tibet-Obs that consists of three regional-scale SM monitoring networks, i.e. the Maqu, Naqu, and Ngari (including Ali and Shiquanhe) networks. This surface SM dataset includes the original 15 min in situ measurements collected by multiple SM monitoring sites of the three networks and the spatially upscaled SM records produced for the Maqu and Shiquanhe networks. Comparisons between four spatial upscaling methods – i.e. arithmetic averaging, Voronoi diagrams, time stability, and apparent thermal inertia – show that the arithmetic average of the monitoring sites with long-term (i.e. ≥ 6-year) continuous measurements is found to be most suitable to produce the upscaled SM records. Trend analysis of the 10-year upscaled SM records indicates that the Shiquanhe network in the western part of the TP is getting wet, while there is no significant trend found for the Maqu network in the east. To further demonstrate the uniqueness of the upscaled SM records in validating existing SM products for a long-term period (∼10 years), the reliability of three reanalysis datasets is evaluated for the Maqu and Shiquanhe networks. It is found that current model-based SM products still show deficiencies in representing the measured SM dynamics in the Tibetan grassland (i.e. Maqu) and desert ecosystems (i.e. Shiquanhe). The dataset would also be valuable for calibrating/validating long-term satellite-based SM products, evaluation of SM upscaling methods, development of data fusion methods, and quantifying the coupling of SM and precipitation at a 10-year scale. The dataset is available in the 4TU.ResearchData repository at https://doi.org/10.4121/12763700.v7 (Zhang et al., 2020).