The freezing index (FI) is an important index used in investigations of climate change, frozen ground degradation and frost heave resistance engineering design. In view of the fact that the deterministic effects of latitude and elevation are not considered in the frequency calculation of FI, we proposed an index-freezing method that considers the certainty effects of both elevation and latitude by referring to the index-flood method in this paper. The correlations between the FI and certainty factors (elevation and latitude) were obtained by multiple regression analysis. The effects of latitude and elevation were then removed by nondimensionalisation, and dimensionless FI sequences were subsequently obtained. Finally, the index-freezing method was verified by regional probability analysis. Using the daily average temperature data recorded at 10 major meteorological stations over the 1960-2020 period in Ningxia, the calculation process of the FI and its frequency distribution were provided. The results showed that the proposed FI method can not only remove the certainty effects of elevation and latitude but can also consider the uncertainty associated with interannual FI variations, thus providing more scientific, reasonable and accurate results. The generalised extreme value (GEV) distribution is the optimal frequency distribution of the nondimensional regional FI. The estimation errors of the missing data tests were mostly within 10%, and the residual sum of squares (RSS) and root-mean-square error (RMSE) values were also lower than those obtained through spatial interpolation, thus indicating that the interpolation preci-sion of the proposed FI method was optimal.

来源平台：COLD REGIONS SCIENCE AND TECHNOLOGY