The evolution of the average freezing depth and maximum freezing depth of seasonal frozen soil and their correlations with the average winter half-year temperature in Heilongjiang Province are analyzed. Linear regression, the Mann–Kendall test, and kriging interpolation are applied to freezing depth data from 20 observation stations in Heilongjiang Province from 1972 to 2016 and daily average temperature data from 34 national meteorological stations collected in the winters of 1972–2020. The results show that the average freezing depth decreases at a rate of 4.8 cm/10 a and that the maximum freezing depth decreases at a rate of 10.1 cm/10 a. The winter half-year average temperature generally shows a fluctuating upward trend in Heilongjiang Province, increasing at a rate of 0.3°C/10 a. The correlations between the average and maximum freezing depths and the winter half-year average temperature are -0.53 and -0.49, respectively. For every 1°C increase in the average temperature during the winter half of the year, the average freezing depth decreases by 3.85 cm, and the maximum freezing depth decreases by 7.84 cm. The average freezing depth sequence mutated in 1987, and the maximum freezing depth sequence mutated in 1988. The average temperature in the winter half-year displayed multiple abrupt changes from 1972 to 2020. The spatial variations in the average and maximum freezing depths are basically consistent with those in the average winter half-year temperature. These research results provide a theoretical basis for the design and site selection of hydraulic structures in cold areas and for regional development and agricultural planning.

Freeze–thaw cycles cause serious soil erosion, which makes the prevention, control and management of solonetzic lands in the Songnen Plain challenging. The use of soil-aggregate-promoter (SAP) is highly favoured because of its energy-saving and efficient characteristics; however, SAP is rarely used in the improvement of solonetzic soil in cold regions. To fill this gap, we studied the effects of different experimental conditions on the physicochemical properties of solonetzes; the investigated conditions included the number of laboratory-based freeze–thaw cycles (with 0, 1, 3, and 5 cycles), initial moisture content (0%, 18%, 24%, and 30%) and SAP application amount (0 g/m2, 0.75 g/m2, 1.125 g/m2, and 1.5 g/m2). The results showed the following: (1) The soil pH value decreased significantly as the SAP application rate increased, and the effect of the initial moisture content and number of freeze–thaw cycles on soil pH was not significant. (2) SAP effectively reduced the soil electrical conductivity (EC), but a certain threshold was apparent, and the factors studied had significant effects on EC. (3) SAP effectively optimised the soil macroaggregates content and inhibited the damage posed by freeze–thaw cycles to the soil structure. These results provide an important theoretical basis for the effective prevention and control of solonetzes in the Songnen Plain of Northeast China.