Investigation of spatial and temporal variability of river ice phenology and thickness across Songhua River Basin, northeast China

The regional role and trends of freshwater ice are critical factors for aquatic ecosystems, climate variability, and human activities. The ice regime has been scarcely investigated in the Songhua River Basin of northeast China. Using daily ice records of 156 hydrological stations across the region, we examined the spatial variability in the river ice phenology and river ice thickness from 2010 to 2015 and explored the role of snow depth and air temperature on the ice thickness. The river ice phenology showed a latitudinal distribution and a changing direction from southeast to northwest. We identified two spatial clusters based on Moran's I spatial autocorrelation, and results showed that the completely frozen duration with high values clustered in the Xiao Hinggan Range and that with low values clustered in the Changbai Mountains at the 95 % confidence level. The maximum ice thickness over 125 cm was distributed along the ridge of the Da Hinggan Range and Changbai Mountains, and the maximum ice thickness occurred most often in February and March. In three subbasins of the Songhua River Basin, we developed six Bayesian regression models to predict ice thickness from air temperature and snow depth. The goodness of the fit (R2) for these regression models ranged from 0.80 to 0.95, and the root mean square errors ranged from 0.08 to 0.18 m. Results showed significant and positive correlations between snow cover and ice thickness when freshwater was completely frozen. Ice thickness was influenced by the cumulative air temperature of freezing through the heat loss of ice formation and decay instead of just air temperature.