Although water temperature is one of the most important factors influencing hydrochemistry and river ecology, long-term monitoring and modelling of stream thermal temporal variability are uncommon. There is sparse research regarding the thermal regimes of Arctic rivers, especially in Svalbard, a geographical hotspot affected by extreme climate change and Arctic amplification. There is a need for improvement and better understanding of the factors influencing the stream water temperature regime. To address this research gap, we present a study of the non-glaciated arctic catchment, Fuglebekken (Spitsbergen, Svalbard). We propose methods for reconstructing the thermal regime of the Arctic stream based on available in-situ data. This study evaluates different sets of input variables with hourly time steps required to explain the variability in water temperature. The study comprises two modelling approaches, a stochastic transfer function Multiple Input Single Output and a supervised machine learning technique, Gaussian Process Regression, to simulate the water temperature in the years 2005-2022. The ground temperature at a depth of 20 cm and total solar radiation were found to be the main forcings that explain most of the water temperature variability. The outputs of both models showed similar tendencies and patterns. A diurnal warming trend of 0.5-3.5 °C per decade has been detected in stream water temperature throughout the summer season. The highest increase of 6.0 °C in the water temperature in 2005-2022 was found to be present in the second part of June. The outcomes prove that the thermal regime of the Fuglebekken stream is sensitive to ongoing climatic changes. This variability is an important factor with many environmental implications.
Keywords: Arctic catchment; Polar hydrology; Svalbard; Water temperature modelling.
Copyright © 2024 The Authors. Published by Elsevier B.V. All rights reserved.