The study area lies in East Africa, specifically in the latitude range of 1˚ to 12˚S and the longitude range of 29˚ to 41˚E (Figure 1). The nation is bordered to the north by Kenya and Uganda, to the west by Burundi, Rwanda, and the Democratic Republic of the Congo, to the south-west by Malawi and Zambia, to the south by Mozambique, and to the east by the Indian Ocean. The region’s climatic diversity is mostly due to the nation’s intricate topographical terrain. Bimodal rainfall patterns occur in the north, northeastern highlands, Pemba and Unguja on the Island of Zanzibar, the center, and the southern coast (Luhunga et al., 2016) . Unimodal rainfall patterns occur in the western and southwestern highlands.

The shifting of the Inter-Tropical Convergence Zone (ITCZ) is the primary cause of various patterns of rainfall (Borhara et al., 2020) . This zone runs through Tanzania from October to February, then the reverse direction from March to May. Tanzania’s seasonal rainfall varies greatly by region; in the wettest months, some regions receive as much as 300 mm of rain every month. Tanzania experiences monthly rainfall ranging from 50 to 200 mm on average.

Tanzania’s average yearly temperature varies by region by 14.4˚C and 1837 mm of rainfall (Luhunga et al., 2016; Borhara et al., 2020) . In comparison to other locations, the western and coastal regions have milder temperatures. In contrast, May marks the start of the low-temperature season, which lasts until August or September. The season with the maximum temperatures in all of the regions starts in October and lasts until February or March. The regions’ lowest and highest average annual temperatures, respectively, range from 9.6˚C - 22˚C to 19.1˚C - 30.7˚C.

Since it was not possible to have daily observed minimum and maximum temperature data from the 29 stations (Table 1) from the study area, the climate data were obtained from National Aeronautical and Space Administration (NASA) retrieved from https://power.larc.nasa.gov/data-access-viewer/. For many years, the National Aeronautics and Space Administration (NASA) has funded satellite systems and studies that provide crucial data for the study of climate and climate processes through its Earth Science research program. These contain estimations of meteorological variables and surface solar energy fluxes that have been long-term averaged climatologically. The time series format further contains

the mean daily values of the underlying solar and meteorological data. In places where surface measurements are scarce or nonexistent, these satellite and model-based products have proven to be accurate enough to deliver trustworthy data on solar and meteorological resources.

An analysis of earlier research using NASA/POWER records and data from the US Cooperative Observer Program (COOP) revealed strong agreement between the two datasets (White et al., 2008) . NASA/POWER climate datasets have also been verified by Van Wart et al. (2015) , White et al. (2008) , and Bai et al., (2010) . As a result, a database containing past climatic information for every station was created. NASA climate data characteristics are accurate and agreeable with a spatial resolution of 1˚ latitude and 1˚ longitude with observed climate data which are archived in the Agro-climatology Archive (Chandler et al., 2013) .

The results of the trend analysis from 29 stations time series of annual averages of extreme temperature indices during 1982-2022 are being summarized in Table 4. The frequency of the percentages of eight temperature indices at 29 analyzed stations are grouped based on the type of trend employed as; significant positive trend 29.3%, non-significant positive trend 37.9%, significant negative trend 13.4%, non-significant negative trend 19.4% and no trend 0% as summarized in Table 3.

The maximum value of daily maximum temperature (TXx) and the minimum value of maximum temperature (TXn) present positive average trend (increasing) but the trends are not statistically significant as in Figure 3(a), Figure 3(b) respectively. The maximum value of daily minimum temperature (TNx) and the minimum value of daily minimum temperature (TNn) indicate positive average trend (increasing) and they are statistically significant (Figure 3(c), Figure 3(d)).

During the period of analysis the annual TXx had an increasing rate of 0.005˚C/ year in Tanzania. From Table 3, only 18.79% of the series had a statistical significant trend in which 10.34% were positive and 3.45% negative. Generally, the

increasing trends were predominant in 68.96% of the series. The increasing pattern was noticed clearly in the years from 1997-2006. The increase of TXx index was concentrated mainly in 17 regions in Eastern, Northeastern, south highlands, western and Lake Victoria areas. The statistical significant trends were observed in the regions of Mahenge, Mtwara and Same (Table 4, Figure 2(a)). Only one station in the Northern part, which is Arusha, had a significant decrease spatial pattern (Table 4, Figure 2(a)). The regions of Mwanza, Geita, Shinyanga, Tabora, Singida, Dodoma, Moshi and Mbeya had shown decreasing trends but not statistically significant as shown in Figure 2(a).

In the analysis it had been revealed that the annual TXn had an increasing rate of 0.009˚C/year. In Table 3, only 17.24% of the series had a positive statistical significant trend. In general, the increasing trends were predominant in 72.41% of the series. The increasing pattern had been seen in the years between 1998 and 2017 as shown in Figure 3(b). The increasing spatial pattern dominance was

Note: *Statistical Significant at confidence level of 0.05.

seen in all areas of Lake Victoria, Western area and southern highlands as in Figure 2(b). Only regions of Tanga, South Pemba, North Pemba, Zanzibar and Dar es salaam had a significant increase (Table 4, Figure 2(b)). The decrease was seen in Arusha, Moshi, Same, Morogoro, Mahenge, Singida, Dodoma and Mtwara but not statistically significant (Figure 2(b)).

The annual maximum value of daily minimum temperature (TNx) had an increase rate of 0.01˚C/year. In Table 3, only 48.28% of the series had a positive

statistical significant trend. In general, the increasing trends were predominant in 86.21% of the series. The increasing pattern had been seen in the years between 1998 to 2017 as in Figure 3(c). The spatial pattern significant increase had been seen in 14 regions among the 29 stations which include Musoma, Kigoma, Mwanza, Moshi, Same, Singida, Dodoma, Morogoro, Mahenge, Iringa, Njombe, Mbeya, Lindi and Mtwara (Table 4, Figure 2(c)). The increasing pattern but not statistically significant were observed in the regions of Bukoba, Simiyu, Shinyanga, Manyara, Arusha, Tanga, South Pemba, North Pemba, Sumbawanga, Songea and Dar es salaam. In the Figure 2(c) the decrease had been noticed in the regions of Geita, Tabora, Katavi and Zanzibar with no statistical significant.

The annual minimum value of daily minimum temperature (TNn) had an increase rate of 0.02˚C/year. In Table 3, only 37.93% of the series had a statistical significant trend in which positive statistical trend was 34.48% and 3.45% was for negative statistical trend. In general, the increasing trends were predominant in 86.21% of the series. The temporal increasing pattern was seen in the years of 1982 to 1984, 1997 to 2003, 2005 to 2006 and 2009 to 2015 in Figure 3(d). The spatial pattern significant increase had been seen in the regions of Kigoma, Simiyu, Babati, Tanga, South Pemba, North Pemba, Iringa, Mbeya, Njombe and Songea (Table 4, Figure 2(d)). The significant decrease was only observed in Bukoba area (Table 4, Figure 2(d)). The increasing pattern but not statistically significant was seen in the regions of Musoma, Mwanza, Geita, Shinyanga, Dodoma, Arusha, Moshi, Same, Singida, Morogoro, Mahenge, Katavi, Sumbawanga, Mtwara and Dar es salaam. The decrease was seen in Tabora, Zanzibar and Lindi regions with no statistical significant.

The percentage of days when daily maximum temperature > 90^th percentile in a year (TX90p) and the percentage of days when daily maximum temperature < 10^th percentile in a year (TX10p) show positive (increasing) and negative (decreasing) trends but they are not statistically significant respectively. The percentage of days when daily minimum temperature > 90 percentile in a year (TN90p) and the percentage of days when daily minimum temperature < 10^th percentile in a year (TN10p) show positive (increasing) and negative (decreasing) trends and they are statistically significant as indicated in Figures 5(a)-(d) respectively.

During the period of analysis the annual TX90p had an increase of 0.103 days/year in Tanzania. From Table 3, only 41.38% had significant positive trends in which there was no any negative significant trend. Generally, the increasing trends were predominant in 65.52% of the series. The temporal increasing trends pattern was seen in the years of 1983, 1988, 1991, 1998 to 2003, 2005 to 2006, 2009 to 2011, 2013, 2015 to 2017 and 2019 as seen in Figure 5(a). The spatial pattern of the significant increasing had been seen in the regions of Moshi, Same, Tanga, South Pemba, North Pemba, Zanzibar and Dar es salaam (Table 4, Figure 4(e)). The increasing trend but not statistically significant was observed in the regions of Geita, Shinyanga, Manyara, Arusha, Sumbawanga, Mbeya, Iringa, Njombe, Mahenge, Lindi and Mtwara. The decreasing trend was seen in Bukoba, Musoma, Mwanza, Simiyu, Tabora, Katavi, Singida, Dodoma, Morogoro and Songea with no statistical significant as shown in Figure 4(e).

In the analysis the annual TX10p had a decrease rate of 0.007 days/year in Tanzania. From the Table 3, only 37.93% had significant trends in which 20.69% had positive and 17.24% had negative significant trends. In general the increasing trends were predominant in 58.62% of the series. The temporal increasing trends pattern was observed in the years of 1984 to 1986, 1989 to 1990, 1990, 1996 to 1997, 1999, 2006, 2008, 2018, 2020 to 2022 as depicted in Figure 5(b). The spatial pattern of the significant increasing had been seen in the regions of Musoma, Simiyu, Shinyanga, Tabora, Singida and Morogoro (Table 4, Figure 4(f)). In the meantime, the significant decreasing trends were seen in the regions of Tanga, South Pemba, North Pemba, Zanzibar and Dar es salaam (Table 4, Figure 4(f)). The increasing trends were observed in the regions of Geita, Mwanza, Arusha, Moshi, Same, Manyara, Dodoma, Mahenge, Songea and Mtwara but were not statistically significant. The decreasing trend which is not statistically significant was observed in the regions of Bukoba, Kigoma, Sumbawanga, Mbeya, Njombe, Iringa and Lindi as referred to Figure 4(f).

During the analysis the annual TN90p had an increasing rate of 0.309 days/year in Tanzania. Only 79.31% had positive significant trends as shown in Table 3.

Generally, the increasing trends were predominant in 96.55% of the series. The fluctuations were also seen in the time series in which the temporal pattern trends were observed in the years of 1982 to 1983, 1987 to 1988, 1991, 1998, 2003, 2005 to 2007, 2009 to 2011, 2015 to 2017 and 2019 as depicted in Figure 5(c). The spatial pattern significant increasing trends were seen in the regions of Kigoma, Mwanza, Musoma, Simiyu, Arusha, Moshi, Same, Babati, Singida, Dodoma, Morogoro, Mahenge, Iringa, Mbeya, Njombe, Songea, Mtwara, Lindi, Zanzibar, South Pemba, North Pemba, Tanga and Dar es salaam (Table 4, Figure 4(g)). The decreasing trends but not statistically significant was seen only in Katavi region. The increasing trends with no statistical significant were observed in the regions of Bukoba, Geita, Shinyanga, Tabora and Sumbawanga as referred in Figure 4(g).

The analysis of annual TN10p had a decreasing rate of 0.256 days/year. Only 82.76% had negative significant trends. Generally, the decreasing trends were predominant in 96.55% of the series. The fluctuations were also seen in the time series in which the temporal pattern trends were observed in the years of 1982 to

1983, 1987 to 1988, 1990 to 1991, 1994, 1997 to 1998, 2000 to 2007, 2009 to 2011, 2013 to 2015 and 2017 to 2019 as shown in the Figure 5(d). The spatial pattern significant decreasing trends were observed in the regions of Musoma, Mwanza, Simiyu, Kigoma, Singida, Dodoma, Babati, Arusha, Moshi, Same, Tanga, South Pemba, North Pemba, Zanzibar, Dar es salaam, Morogoro, Mahenge, Iringa, Njombe, Mbeya, Sumbawanga, Songea, Lindi and Mtwara (Table 4, Figure 4(h)). The decreasing pattern with no statistical significant was seen in Geita, Shinyanga, Tabora and Katavi. The only increasing spatial pattern with no statistical significant was observed in Bukoba as shown in the Figure 4(h).

In comparison, there has been an increase in the percentage of warm nights (TN90p) and warm days (TX90p). The trend in the percentage of TN90p is more prominent in larger regions of the country, suggesting that the nights are warming more quickly, as illustrated in Figure 4(g). In the meantime, the results obtained are in line with other research findings in the study area (Chang’a et al., 2017; Chang’a et al., 2021) . They are also in line with the observed temperature extreme trends in various parts of the world as shown in the IPCC Reports (IPCC, 2007; Chang’a et al., 2017; Wang et al., 2018) , which show that the number of warm days (TX90p) and nights (TN90p) are increasing. Seemingly, the number of cold days (TX10p) and nights (TN10p) has been decreasing. On the extreme absolute temperature indices, the findings show that the warming nights have increased and the coldness have decreased comparing with the warm days and cold days. Similar phenomena agree with various obtained results from different parts of the world.