1. Basic data preparation: (1) Using the daily temperature data of the quality control station, calculate the temperature change at 24 h, 48 h and 72 h, respectively; (2) Calculate the average daily temperature of 57 years from 1961 to 2017, and get the moving average out of the average daily temperature for 5 days. (3) Calculate the daily temperature anomaly value based on the daily maximum temperature data and the multi-year average data.

Temperature-changing data based on the maximum daily temperature. When calculating the daily 24 h temperature change data, the daily maximum temperature of the day is subtracted from the maximum temperature of the previous day, and the difference obtained is recorded as the 24 h temperature change data of the day; the 24 h temperature change data of January 1, 1961 adopted the daily maximum temperature data of December 31, 1960. Similarly, calculate the difference between the daily maximum temperature of the day and that of the previous second day and that of the third day respectively. The temperature change data of 48 h and 72 h on January 1, 1961 used the daily maximum temperature data on December 29, 30, and 31, 1960. We finally get the daily 4 h, 48 h and 72 h temperature change data from January 1, 1961 to December 31, 2017.

The data of the multi-year daily maximum temperature. Take the average value of 57 years as the data of the multi-year daily maximum temperature for each station from January 1, 1961 to December 31, 2017, and calculate the average value from January 1 to December 31 of the 57 years (1961–2017). For the February 29 of the leap year, take the average of the 14-year sample data for 57 years. Then, calculate the daily average value from January 1 to December 31 again with the 5-day moving average of the two days before and after, and finally get the multi-year average data of 57 years from January 1 to December 31.

The data of daily maximum temperature anomaly. Using the daily maximum temperature data from January 1, 1961 to December 31, 2017, subtract the multi-year average of the day, and get the data of the daily maximum temperature anomaly from January 1, 1961 to December 31, 2017.

2. Identification of the temperature-rising process: (1) Based on the definition of the temperature-rising process shown in Table 2, use the daily 24 h temperature change data to identify the starting and ending days of the temperature-rising process; (2) combine the daily maximum temperature and its anomaly, as well as the daily 24 h, 48 h and 72 h temperature change values to determine the maximum temperature rises of the process during the temperature-rising process of 24 h, 48 h and 72 h, and obtain the maximum value of the process extreme maximum temperature and process daily maximum temperature anomaly range. Then we get the values of the various elements of the successive temperature-rising processes to form the dataset file of the temperature-rising process for each station.

Various elements of one temperature-rising process are composed of the specific date of the start and end of the temperature-rising process described above, and the number of process durations, the identified process temperature rise, the maximum 24 h, 48 h, 72 h temperature rise, as well as the maximum anomaly range and extreme maximum in the process. We sorted out the above-mentioned elements of all temperature-rising processes from January 1, 1961 to December 31, 2017 collected by the stations to form the temperature-increasing process dataset.

2.2.3 The calculation of the extreme temperature-rising process

The calculation of single-element strength. Based on the dataset of the temperature-rising process of the station, we analyzed the normalized values of the following four elements: the process temperature-rising range, the maximum 24 h temperature rise, the process extreme maximum temperature and the process daily maximum temperature anomaly range. The four single-element indices could quantitatively characterize the temperature rise process from different angles.

Comprehensive strength index calculation based on the four elements. Among the four single-element strength indices of each process, the index of the process temperature-rising range (*IZFD*) reflects the overall strength of the temperature-rising process; the index of the process maximum 24 h temperature-rising range (*IZFD*24) reflects the short-term temperature-rising strength during the process; the index of the process extreme maximum temperature (*IZTG*) reflects the absolute maximum temperature strength during the temperature-rising process; the index of the process daily maximum temperature anomaly range (*IZJP*) reflects the deviation of the daily maximum temperature from the average of the same period in the temperature-rising process over the years. The sum of the equal weights of the above four single-element strength feature normalization indices is used as the comprehensive evaluation index *IZ4* of the single-station temperature-rising process, and the calculation formula is:

*IZ4＝IZFD＋IZFD* 24*＋IZTG＋IZJP * (1)

According to the comprehensive index *IZ*4 from high to low, the first 5% of the temperature-rising processes would be selected to form the dataset of the extreme temperature-rising process.