A comprehensive analysis of three meteorological datasets for the East Siberian continuous permafrost zone: long-term changes in air temperature, snow depth and precipitation

Le résumé fourni par la source

Abstract. Climate and its evolution strongly influence numerous natural processes, and in permafrost areas this impact outweighs that of any other factors. The main publicly available databases consist of data only for 167 meteorological stations within Republic of Sakha (3,08 ml km2), Magadan Obslat (0,46 ml km2) and Chukotka (0,72 ml km2), which makes the meteorological network in that region – almost entirely underlain by continuous permafrost – one of the sparsest on the planet, at approximately 1 station per 25000 km2. The longest series dates back to 1834 at the first official meteorological station in Yakutsk-24959, while the shortest to 1987 in Anabar-21608), however, data availability is highly inconsistent, and more than half of meteorological stations contain gaps of up to 40 % in their timeseries. Despite the abundance of reanalysis products, here we aim to assess the current meteorological network, the available observation data, and to merge three different datasets into a unified database containing the following parameters: mean, maximum, and minimum daily air temperature, daily precipitation totals and snow cover thickness. The resulting database spans the period from 1 January 1966 to 30 September 2025, with daily temporal resolution, and includes source references as well as the number of measurements recorded for each day as a quality indicator parameter. We carried out automated cleaning for obvious mistakes and typos as well as deleting duplicates. Chosen parameters are particularly important for a wide range of researchers, as well as for society and policymakers. This dataset has been used for filling gaps in air temperature time series using machine learning approaches. For convenient access, we have developed an online dashboard that presents the main meteorological parameters at daily, decadal, monthly, and annual scales. The database presented and described in this article is available for download at https://doi.org/10.5281/zenodo.19883706 and at https://robertseesaw.shinyapps.io/meteo-dashboard/.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.