An improved method for calculating the freezing/thawing index using monthly and annual temperature data

Abstract

AbstractChanges in soil thermal regimes in cold climates have widespread impacts on hydrology, ecology, and the carbon cycle. The annual freezing and thawing index, which is generally calculated using daily temperature, has been widely used to estimate the freezing depth, active layer thickness, and the distribution of permafrost. However, continuous and reliable daily temperature data are scarce in cold climates, while monthly and annual temperature data are more readily available. If daily temperature data are unavailable, these indices can be estimated based on monthly or annual temperature data. In this study, we developed a resampling method for estimating the annual freezing and thawing index and compared the results with those produced by the existing methods. Daily temperature data with a 0.5° resolution over the Northern Hemisphere during 1901–2012 were used to calculate the freezing/thawing index, and then the monthly and annual temperature were calculated and three different approaches were used to estimate the daily temperature and the freezing/thawing index. When the monthly data were used, the resampling method produced the smallest relative error (RE) and mean bias error (MBE), and the largest correlation in estimating the two indices, compared to the two other methods. Although the annual temperature data usually underestimate the freezing/thawing index, the RE is still <5% over most of the high‐latitude regions. The results suggest that if the daily temperature can be reliably estimated using the resampling method, the thermal regimes of permafrost can be reliably estimated using modelled monthly temperature and/or reconstructed past monthly/annual temperature. These estimations can also be used to validate modelled paleo‐permafrost and its variations. Additionally, our results indicate that after the 1970s the annual freezing index (DDF) increased substantially, while the frost index (FI) decreased substantially.