Every single observation in a series is subjected to quality control (QC) protocols. For both the blended and non-blended station series, these processes are carried out independently. There are three QC flags in use at the moment: Flag = 0 means “valid”, Flag = 1 means “suspect”, and Flag = 9 means “missing”. The following requirements must be met by each component of the daily minimum temperature (TN) and daily maximum temperature (TX): it must be greater

than −90.0˚C and less than 60.0˚C; it must also surpass or equal the daily minimum or maximum temperature (if it exists), it must be less than the long-term average daily maximum temperature for that calendar day plus five times the standard deviation (calculated for a 5-day window centered on each calendar day over the entire period), it must exceed or be equal to the daily mean temperature (if it exists), it cannot be repetitive (i.e., exactly the same) for five consecutive days, and it must exceed the long-term average daily minimum or maximum temperature for that calendar day minus five times the standard deviation (calculated for a 5-day window centered on each calendar day over the entire period) (Project Team ECA & D, 2021) .

In order to quantitatively define temperature extremes in the June to August season over Tanzania, the Expert Team on Climate Change Detection and Indices (ETCCDI) list of 27 temperature and precipitation indices in which 16 indices for temperature and 11 for precipitation, only 8 temperature indices were selected as shown in Table 2. The selected indices are apt for explicit climatic conditions found in the study area. The public can view these extreme temperature and precipitation indices from the ETCCDI at http://etccdi.pacificclimate.org/indices.shtml. According to Feng et al. (2018) , these indices can be divided into two types depending on their indicators: absolute indicators, including minimum Tmin (TNn), maximum Tmin (TNx), minimum Tmax (TXn), and maximum Tmax (TXx). TN10p, TN90p, TX10p, and TX90p are the percentile-based indicators that show how frequently chilly nights, warm nights, and warm days occur.

The study comprises a seasonal trend analysis conducted annually from June to August. The European Climate Assessment & Dataset project (ECA&D) is utilized to calculate climatic indices in accordance with the guidelines provided by ETCCDI. Every index and aggregate period has its trend calculated; in order to perform a trend analysis, at least 70% of the values within a period must contain valid index data, with no missing data (Project Team ECA & D, 2021) . The method was also employed in the work of (Ndabagenga et al., 2023) .

In the current work, the Mann-Kendall (MK) test (Mann, 1945; Kendall, 1975) was taken to assess the presence of trends in extreme climatic indices. This non-parametric test relies on relative rankings within a specified time range to determine the existence of a trend. The Mann-Kendall test statistic is computed as follows:

S = ∑ i = 1 n ∑ j = i + 1 n s i g n ( x j − x i ) , s i g n ( x j − x i ) { 1 i f x j − x i > 0 0 i f x j − x i = 0 − 1 i f x j − x i < 0 } (1)

A positive S value signifies an ascending trend, whereas a negative value denotes a descending trend. The Z value is determined by calculating the variance of the temperature. The computation of the variance (S) is as follows:

var ( s ) = 1 18 [ n ( n − 1 ) ( 2 n + 5 ) − ∑ i = 1 m t i ( t i − 1 ) ( 2 t i + 5 ) ] (2)

A tied group (m) refers to a collection of temperature data points sharing the same value when the sample size is greater than 10 (n > 10) and the t_i is the number of data points in the i^th tied group. The standard Z test statistic is determined using the following equation:

z = S ± 1 Var ( s ) 1 / 2 (3)

This equation takes S − 1 if S > 0, S + 1 if S < 0, and Z is 0 if S = 0. A positive value of Z signifies an upward trend. Otherwise, it shows a descending trend.

The Sen’s slope (Sen, 1968) estimator was utilized to find the extent of the trends in temperature extremes. This method assumes a linear trend, providing a quantification of the temporal change. Sen’s Slope offers an advantage over linear regression as it is unaffected by the presence of outliers and data errors. The equation for Sen’s Slope, considering N data sample pairs, is expressed as follows:

Q i = ( x j − x i ) j − i , i = 1 , 2 , ⋯ , N (4)

The data values at times j and i (where j > i) are represented by x_j and x_i, respectively. In a time series with n values of x_j, the total number of slope estimates N is given by N = n(n − 1)/2.

The eastern, northeastern, southwestern highlands, western portion, places surrounding Lake Victoria Basin, northern part, and southern parts have all shown a considerable increase in extreme warm day trends, as illustrated in Figure 2(a), at a rate of 0.016˚C to 0.048˚C. The statistically significant rise has been observed at a rate of 0.02˚C - 0.04˚C in the eastern, northeastern, and northern regions in the analysis of exceptionally warm nights (Figure 2(c)). Significant drops between 0.02˚C and 0.06˚C has been seen in the western, northwest, and southern regions.

The extreme cold days trends have also been observed in the country in which the significant increase is depicted at the rate of 0.00˚C - 0.08˚C in eastern, northeastern, southwest highlands, southern part, western areas and around Lake Victoria Basin. The statistically significant decrease has been seen in few areas of Sumbawanga region at the rate of 0.04˚C - 0.12˚C (Figure 2(b)). In the meantime, the extreme cold nights have been observed with a statistically significant increase in the rate of 0.045˚C - 0.075˚C in the areas of southwestern highlands, eastern parts, Lake Victoria Basin in Bukoba region and Musoma region. In the central, western and southern parts increase at the rate of 0.030˚C - 0.045˚C and they are statistically significant. The statistically significant decrease is seen in the areas of Babati, Arusha and Moshi at the rate of 0.015˚C - 0.030˚C (Figure 2(d)). Generally, the analysis to this absolute extreme temperature during JJA season there is an increase of coldness during the day and nights compared with hotness during days and nights.

Figure 3(a)-(d) depicts the temporal changes in the regionally averaged June-August season for each year of severe temperature events in Tanzania from 1982 to 2022. Figure 3(a) shows a substantial increase in extreme warm days (TXx) at a rate of 0.01677˚C per year during the season, while Figure 3(c) shows no statistically significant increase in extreme warm nights (TXn) at a rate of 0.014˚C per year during the season. The extreme cold days (TNx) increased at the rate of 0.00594˚C in the season every year with no statistically significant (Figure 3(b)). In the meantime, the extreme cold nights (TNn) increased at the rate of 0.00973˚C in the season every year with no statistically significant (Figure 3(d)). Generally, in all absolute indices there is an interannual variation during the season of JJA, and in particularly the extreme warm indices which shows consistent increase from the year 2002 to 2022.

During the analysis of these indices, it was disclosed that the warm days (TX90p), had significantly increased in the eastern, northeastern, southern parts, western areas and around Lake Victoria Basin at a rate of 0.125 - 0.150 days in a

season in each year. In the areas of central, some parts of Lake Victoria Basin, southwestern highlands, few areas of western country and some northern parts had a statistically significant increase at the rate of 0.050 - 0.125 days in a season in each year (Figure 4(a)).

The warm nights (TN90p) had a statistically significant increase at the rate of 0.09 - 0.15 days in a season in each year in the eastern, northeastern parts, and Musoma region. In the areas of central, some parts of Lake Victoria Basin and few areas of southwestern parts had a statistically significant increase at the rate of 0.06 - 0.09 days in a season in each year. Slight significant increase has been seen in some few areas of Southern parts especially in Songea region, southwestern highlands and the region of Bukoba, Singida and Shinyanga at the rate of 0.03 - 0.09 days in a season in each year. A statistically significant decrease has been observed in the western parts of the country at the rate of 0.03 - 0.6 days in a season in each year (Figure 4(c)).

According to the data, the number of cold days has decreased nationwide at a pace of 0.04 to 0.12 days per season on an annual basis. A statistically significant increase at the rate of 0.04 to 0.16 days in a season per year has only been observed in a few numbers of places of the southwestern highlands, mainly in the

Mbeya region and the western parts of the Sumbawanga region (Figure 4(b)). In most parts of the country, the number of cold nights has significantly decreased throughout the June-August season. At a rate of 0.03 to 0.15 days per season every year, a statistically significant decline has been observed in the eastern, northeastern, Lake Victoria Basin, southern, southwestern highlands, and central regions of the country (Figure 4(d)).

Figure 5(a)-(d) depicts the temporal variations in Tanzania’s regionally averaged June–August season for each year of severe temperature events from 1982 to 2022. The warm days (TX90p) in Figure 5(a) exhibit interannual fluctuation and a noteworthy rise at a rate of 0.1043 days per season from 1982 to 2022. There is a little increase in warm days from 1982 to 2001, but there is an upward trend from 2002 to 2022. The number of warm nights (TN90p) is also rising significantly, at a pace of 0.0644 days in a season in each year. From 2000 to 2022, the trend shows increasing values, with the small value observed from 1983 to 1999 (Figure 5(c)). The variation of cold days in a season observed to decrease significantly at the rate of -0.0335 days in a season in each year and the value was higher from 1982 to 2000 and thereafter the value was slightly small (Figure 5(b)). Likewise, cold nights (TN10p) appeared to decrease significantly at the rate of −0.0749 days in a season in each year. The value was higher from 1982 to 2000 and then the decrease had small value to 2022 (Figure 5(d)). It is found in this result that the warm days and nights are increasing in which the value of increase in warm days (0.1043 days in a season in each year) are larger than warm nights (0.0644 days in a season in each year) while the decrease rate of cold days (−0.0335 days in a season in each year) was smaller than cold nights (−0.0749 days in a season in each year) which shows that the night cooling was higher than daytime cooling. The overview from the results, it indicates that during JJA season extreme events related to extreme temperatures in the country both significantly increased and decreased with respect to their related indices. The results also demonstrate that the extreme temperature indices during the JJA season have increased significantly in line with the global climate warming expectations.

In comparison, the findings of this study indicate a noticeable increase in warm days and nights, accompanied by a decreasing trend in cold days and nights in the season. Some of the factors that may have contributed to the current seasonal changes include rising greenhouse gas concentrations, which have resulted in overall warming trends, as well as urban settlement development, where human activities and the built environment have occasioned in high temperatures. Moreover, changes in land cover and use practices, such as deforestation and agricultural expansion, have the potential to influence local temperature patterns. Furthermore, the study suggests that anthropogenic influences may contribute to the warming trend observed in extreme daily minimum and

maximum temperatures globally (IPCC, 2012) , with Tanzania potentially affected, as revealed in the current research. The overall results of this study reflect patterns observed in various regions worldwide, where warm days and nights are on the rise while cold days and nights are diminishing, as documented by (Hartman et al., 2013) and (Tong et al., 2019) .