Analysing Changes in Rainfall Dynamics: Onset and Precipitation Probability in Tanzania

doi:10.21203/rs.3.rs-5332492/v1

This study investigates the changes in rainfall dynamics in Tanzania, focusing specifically on the onset and precipitation probability of rainy seasons. The research stems from growing concerns about climate variability in East Africa, which has significant implications for agriculture, water resources, and food security. Despite Tanzania's heavy reliance on rain-fed agriculture, there is a lack of understanding regarding how rainfall patterns are shifting, impacting both local farmers and national policies. Utilizing time series analysis and the Instat Climatic tool for the assessment of rainfall onset dates, Mann-Kendall (MK) test, Sen’s slope estimator (Q₂) for onset dates trend analysis and Markov chain model for precipitation probability, we analysed historical rainfall data from 27 weather stations (regions) across Tanzania. Our findings indicate a noticeable shift in the onset of the rainy seasons, with variability in the trends of rainfall onset across different regions. Furthermore, a consistent pattern wherein southern and central regions tend to witness prolonged waiting times for the start of the rainy season and increase in the probability of rain following a rainy day (rr), while the probability following a dry day (rd) remains relatively stable. These results highlight critical adaptations necessary for agriculture and water management strategies in Tanzania. Late rainfall onset could hinder crop growth, particularly for maize and other staple foods, leading to potential food shortages and economic instability if no adaptive measures are implemented. The findings underscore the need for interdisciplinary collaboration to develop robust frameworks that integrate climatic forecasts into agricultural planning. Understanding the shifting dynamics of rainfall is essential not only for Tanzania but also for other regions facing similar climate challenges, emphasizing the interconnectedness of climate change impacts on food systems and livelihoods around the world.

Rainfall dynamics

Precipitation probability

Rainy season

Tanzania

Onset of rainfall

Regional variability

Rainfall is a critical component of the hydrological cycle and plays a vital role in sustaining ecosystems, agriculture, and water supply in Tanzania. The country relies heavily on rain-fed agriculture, which constitutes nearly 4.4% a year of its GDP (World Bank 2020) and is a primary source of livelihood for approximately 75.5% of its population (FAO 2022). However, recent studies indicate that climatic changes are influencing precipitation patterns across the region, leading to increased variability in rainfall onset and cessation (Jury 2024; Nicholson 2017; Palmer et al. 2023). This variability poses significant challenges to agricultural productivity and food security, making it imperative to investigate the trends and potential impacts of changing rainfall patterns.

In recent decades, Tanzania has experienced notable fluctuations in rainfall due to both natural variability and anthropogenic influences. According to the Intergovernmental Panel on Climate Change (IPCC 2021), the East African region is projected to experience increased temperatures and altered precipitation patterns, with potential consequences for agricultural yields, water availability, and overall economic stability. Additionally, changes in rainfall timing can disrupt traditional farming calendars, resulting in reduced crop yields and increased vulnerability among rural communities (Kanga et al. 2024; Kukal and Irmak 2018; Aishwarya 2024). Several studies have documented significant changes in rainfall patterns across Tanzania, noting both temporal and spatial variability. Research byMbigi et al. (2022) found that while the overall amount of rainfall might remain stable, its distribution has become increasingly erratic, causing challenges for agricultural systems reliant on seasonal rains. The onset and cessation of rainfall are critical for agricultural planning and food security.MacLeod (2018) observed a trend of delayed rainfall onset in many regions of Tanzania. This shift has implications for planting schedules, often leading to a mismatch between crop growth stages and water availability. Additionally, late cessation of rains has been documented by Kavishe et al. (2020), which can lead to prolonged growing seasons but may also increase risks of pest infestations and crop diseases. Research shows that the variability of rainfall has increased over the last few decades, impacting agricultural productivity.Kebacho (2021) utilized Heavy Rainfall Events (HRE) to analyse rainfall variability in Tanzania, revealing that farmers are struggling to adapt to these changing patterns. Variability not only affects crop yields but also complicates the decision-making processes for farmers, making it difficult to implement effective agricultural practices.

The implications of changing rainfall patterns for agriculture in Tanzania are dire. Jones et al. (2023) highlighted that smallholder farmers, who are predominantly dependent on rain-fed agriculture, face increased food insecurity due to these climatic changes. The place-based variations in rainfall further exacerbate the situation, as some regions may receive sufficient rainfall while others experience droughts. This disparity leads to unequal benefits across communities, thereby widening the gap in food security. Given the increasing challenges posed by variable rainfall, there is an urgent need for adaptive strategies. Studies such as those by Arndt et al. (2012) and Jones et al. (2023) emphasize the importance of integrating climate information into agricultural planning. Innovative approaches, including the use of drought-resistant crop varieties, improved water management techniques, and education on climate-smart agriculture, are essential for enhancing resilience among farmers (Rowhani et al. 2011). Furthermore, supporting farmers in building local networks for sharing information can improve adaptive capacities (Šūmane et al. 2018).

Weather prediction remains a crucial area of study, especially in the context of climate variability and its implications for ecological and human systems. A fundamental aspect of meteorology is understanding how the conditions of one day influence the conditions of the subsequent days. This study focuses on two key conditional probabilities: the chance of rain given a dry previous day (denoted as (p(rd)) and the chance of rain given that the previous day was rainy (denoted as (p(rr)). Understanding these probabilities is essential not only for meteorologists but also for urban planners, farmers, and policymakers, as it informs decision-making in various sectors affected by weather changes (Yang et al. 2018). Statistical models that relate to these probabilities provide insight into atmospheric behaviour and can help improve forecasting accuracy. Previous studies have shown that antecedent weather conditions significantly influence the likelihood of future precipitation (Koster et al. 2003; Xu et al. 2022). For instance, it has been established that the chance of rainfall tends to be significantly higher following a rainy day than after a dry day, indicating a potential dependency influenced by atmospheric moisture retention and other meteorological factors (Fowler and Wilby 2010).

This research aims to explore these dependencies further across Tanzania, providing a comprehensive analysis of how previous weather conditions affect future precipitation likelihood by employing statistical modelling techniques. While there is substantial research on the impacts of climate change on rainfall patterns, there is still a need for localized studies that examine specific region responses in different agricultural zones of Tanzania. Furthermore, there is a lack of comprehensive analyses combining quantitative data.

The objectives of this research are threefold: to investigate trends in rainfall onset, assess the variability in these patterns over time, and analyse the implications for agricultural practices in Tanzania. By enhancing our understanding of these dynamics, we aim to contribute to the development of adaptive strategies to mitigate the impacts of climate variability on agricultural livelihoods.

The remaining sections of this study are structured as follows: Section 2 describes the study area, datasets, and methods used in the study. Meanwhile, Section 3 presents the main results, section 4 discussions. Finally, the conclusions highlighted in section 5.

2.1 Study Area

Tanzania is located in East Africa (See Fig. 1), bordered by several countries and bodies of water. To the north, it shares borders with Kenya and Uganda, while to the west, it is bordered by Rwanda, Burundi, and the Democratic Republic of the Congo. To the south, Tanzania shares a border with Zambia, Malawi, and Mozambique. Additionally, the Indian Ocean lies to the east, providing the country with a coastline stretching approximately 1,424 kilometres. The geographical coordinates of Tanzania range from about 1°S to 12°S latitude and 29°E to 41°E longitude, placing it centrally situated within the African continent.

Tanzania's topography is diverse, featuring a combination of highlands, plateaus, and lowland areas. The major geographic regions include the central plateau, which is characterized by broad, flat areas interspersed with valleys and hills (Basalirwa et al. 1999). The highest peak in Africa, Mount Kilimanjaro, is located in northeastern Tanzania, with an elevation of approximately 5,895 meters (19,341 feet) above sea level. Other notable highland regions include the Eastern Arc Mountains and the Usambara Mountains. The coastal regions are generally flat, while the interior features more pronounced topographical variations, including the Great Rift Valley, which runs through the country, contributing to the unique landscape (McSweeney, 2009).

Tanzania experiences two main rainy seasons: the long rains (November-April)-Unimodal rainfall regime and the short rains (March-April-May or MAM and October-November-December, or OND)-Bimodal rainfall regime (Mbigi and Xiao 2021; TMA 2024). The distribution and intensity of rainfall across these seasons vary significant due to several climatic factors (Basalirwa et al. 1999; Mapande and Reason 2005). The ITCZ plays a crucial role in determining rainfall patterns in Tanzania. This zone represents an area where the trade winds from the northern and southern hemispheres meet. During the MAM season, the ITCZ shifts northward, bringing moisture-laden winds and resulting in increased rainfall, particularly in the northern and coastal areas of Tanzania (Byrne et al. 2018). The ENSO phenomenon significantly influences Tanzania's climate variability. An El Niño event typically results in above-average rainfall during the MAM season, while La Niña conditions can lead to drier periods. Therefore, the ENSO cycle is a critical factor affecting the seasonal variability of rainfall in both MAM and OND (Dieppois et al. 2021; Indeje et al. 2000; Mason and Goddard 2001). The Indian Ocean Dipole (IOD) is another significant climatic driver. A positive IOD phase typically results in below-average rainfall, particularly during OND, affecting agricultural productivity. Conversely, a negative IOD phase can enhance rainfall during this season, leading to more favourable conditions for crop growth (Endris et al. 2019; Khan et al. 2021; Xiao et al. 2022).Researchers have noted that understanding these climatic factors is essential for predicting rainfall patterns and effectively managing water resources in Tanzania.

[FIG.1]

2.2 Data collection

In-situ data

Data for Tanzania average monthly precipitation was obtained from the Tanzania Meteorological Authority (TMA). The data collected from manual instruments, spans from 1950 to 2020. In Tanzania, the TMA is essential to weather forecasting and climate monitoring. The TMA prioritises data quality control in order to achieve this obligation. In order to guarantee the accuracy and dependability of the data used for weather forecasts, climate assessments, and environmental planning, effective data quality management is crucial. To preserve and improve the quality of its meteorological data, the TMA uses a number of techniques. Standardised data gathering methods are one of the main strategies. The TMA improves the comparability and consistency of the data gathered by guaranteeing that all weather stations follow the same procedures (SOFF, 2021).

Rain Gauge and Satellite observations data

The Tanzania Meteorological Authority (TMA) has established a policy that provides daily rainfall data solely from five meteorological stations for a maximum duration of five years. While this data is crucial for understanding weather patterns and trends in Tanzania, its limited scope may hinder comprehensive analyses and informed decision-making in various sectors, such as agriculture, water management, and disaster preparedness. To address this limitation, this research will explore the use of alternative daily rainfall data sourced from the Climate Hazards Group InfraRed Precipitation with Station Data (CHIRPS) for selected 27 regions (For selected region details see Table 1). CHIRPS offers a long-term dataset from 1981 to 2023 that integrates satellite imagery and ground station data.

CHIRPS provides a high spatial resolution of 0.05 degrees (approximately 5 km) across the globe. This fine resolution enhances the ability to capture localized weather phenomena, making it particularly useful for regional analyses, such as in East Africa, where precipitation can vary significantly over short distances (Funk et al. 2015). CHIRPS generates precipitation data through a combination of satellite-derived estimates and ground station observations. The algorithm integrates various sources of information, including infrared satellite imagery, to interpolate precipitation data over areas without station coverage. The time series is constructed to be temporally homogeneous, providing continuous daily data back to 1981. The technique employed utilizes a convolutional kernel for data smoothing and bias adjustment based on ground data, which helps ensure that the dataset remains consistent over time (Funk et al. 2015). In East Africa, CHIRPS has proven to be a valuable tool for monitoring precipitation variability and extremes. Studies have shown that CHIRPS precipitation estimates correlate strongly with ground measurements, enhancing its applicability for agricultural planning, drought monitoring, and climate change research. For example, a study by Dinku et al. (2018) and Harrison et al. (2022) highlighted that CHIRPS data effectively captures seasonal rainfall patterns, providing a reliable resource for stakeholders in the region. The data available from https://data.chc.ucsb.edu/products/CHIRPS-2.0/ website for downloading.

[Table 1]

2.3 Criteria for the assessment of rainfall Onset and Statistical techniques

Rainfall onset is a critical element in understanding the climatic patterns of East Africa, including Tanzania, it influences agriculture, water resources, and ecological habitats. Tanzania is characterized by unimodal (these areas include Southern (Mtwara, Lindi, and Ruvuma and southern Morogoro), Western ( Tabora, and Kigoma regions), and central parts of the country (Dodoma, and Singida regions) and bimodal (these areas include the northern coast (Tanga, Dar es salaam, north Morogoro, and Coast regions, and Zanzibar Islands), northeastern highlands (Kilimanjaro, Arusha, and Manyara regions), and areas surrounding Lake Victoria (Mara, Kagera, Simiyu, Shinyanga, Geita, and Mwanza) rainfall patterns, with distinct short and long rainy seasons (TMA, 2022,2023).

Rainfall Onset: The date when the first significant rainfall occurs in a given period, marking the beginning of a rainy season. This can be defined using various criteria, such as the amount of rainfall and its duration. For example, an event may be classified as an onset if it exceeds a specified threshold (typically between 5mm and 20mm) within a 24-hour period, which was used in this study. For MAM season earliest onset dates is 1st March and latest possible date is 31 May and for OND season earliest onset dates is 1st October and latest possible is 31 December (Jolliffe and Sarria-dodd 1994; Odenkunle 2004; Segele and Lamb 2005; Shaw 1987). To facilitate this analysis, the Instat Climatic Tool Version 3.036 will be utilized (Stern et al. 2002). This software provides a comprehensive framework for climatic data analysis, incorporating various statistical methods to assess climate variability and trends. Instat statistical software (Interactive Statistics) is designed to assist researchers and practitioners in analysing meteorological data effectively. The tool offers functionalities such as graphical analysis, statistical tests, and climate projections, making it an invaluable resource for those studying climatic influences on agricultural activities. Version 3.036 includes updated features that enhance data processing and improve users' experience in determining critical climatic events, such as the onset of rain.

Detecting trends in time series data is crucial for understanding the dynamics of climate variables such as rainfall onset. In Tanzania, where climate variability has significant implications for agriculture and water resources, applying robust statistical methods to identify trends is imperative. One of the most widely used methods for trend detection in time series data is the Mann-Kendall (MK) test (Hamed 2008; Hamed and Rao 1998; Hu et al. 2020; Yue et al. 2002) .This non-parametric test is particularly effective for assessing trends in time series without requiring the data to conform to any specific distribution (Mann 1945) (see Appendix A).

Moreover, this study will utilize the Sen’s Slope Estimator (also known as m-slope ($\:{\text{Q}}_{2}$)) as a method for detecting the slope or trend in our data (Jiqin et al. 2023). The Sens Slope Estimator is a non-parametric approach that provides a robust estimate of the slope in a time series dataset, particularly useful when dealing with outliers or non-normally distributed data (da Silva et al. 2015). By applying this estimator, I aim to obtain reliable trends that reflect changes over time, enhancing the accuracy of my findings. The Sen’s Slope Estimator is often applied in environmental science, particularly in assessing trends in climate data. One notable example is its use in analysing long-term precipitation or temperature records to detect trends in climate change. In a study by Yue et al. (2003), the Sens Slope Estimator was employed to analyse annual streamflow data across various river basins. The research assessed changes in hydrological patterns and trends, which are critical for water resource management and understanding ecological impacts. Another application can be found in the field of agriculture, where the Sens Slope Estimator is used to analyse crop yield data over time to identify trends related to impacts from climate variability (Jiqin et al. 2023) (see Appendix B).

2.4 Determining Precipitation Probability (Markov chain models)

One effective statistical tool for modelling sequences of rainfall events is the Markov model. This study explores the application of Markov models in predicting the probabilities of rain, providing insights into their theoretical foundations and practical implications. Markov models are stochastic processes that rely on the principle of memory lessness, meaning the future state of a process depends only on its current state, not on its past states (Feller, 1959). In the context of rainfall, a simple two-state Markov model can be employed, wherein the states are defined as "rain" $\:\left({\text{P}}_{\text{r}\text{r}}\right)\:$and "no rain" $\:\left({\text{P}}_{\text{r}\text{d}}\right)$. The transition probabilities between these states can be estimated from historical weather data, allowing for predictions about future rainfall.

The definition of a Markov chain Eq. 1 to 4:

Suppose $\:\left\{{\text{X}}_{\text{n}},\text{n}\ge\:0\right\}$ is a discrete time stochastic process with a countable state space S. It is known as a Markov chain if $\:\forall\:\:\:n\ge\:1$ and $\:\forall\:\:{x}_{0},{x}_{1},\dots\:,{x}_{n-1},i,j\:$in S,

$$\:\text{P}\left[{\text{X}}_{\text{n}+1}=\text{j}\left|{\text{X}}_{\text{n}}=\text{i},{\text{X}}_{\text{n}-1}={\text{x}}_{\text{n}-1,}\dots\:{\text{X}}_{0}={\text{x}}_{0}\right.\right]=\text{P}\left[{\text{X}}_{\text{n}+1}=\text{j}\left|{\text{X}}_{\text{n}}=\left.\text{i}\right],\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\left(1\right)\right.\right.$$

$\:\left[{\text{X}}_{\text{n}-1}={\text{x}}_{\text{n}-1,}\dots\:{\text{X}}_{0=}\left.{\text{x}}_{0}\right]\right.$ is referred to as past, the event $\:\left[{\text{X}}_{\text{n}},=\text{i}\right]$ is referred to as present and the event $\:\left[{\text{X}}_{\text{n}+1}=\text{j}\right]$ is referred to as future.

$$\:\left\{{\text{X}}_{\text{n},}\text{n}\ge\:0\right\},\:\text{i}\text{f}\:{\text{X}}_{\text{n}}=\text{i}\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\left(2\right)$$

The system is in state $\:i$ at time $\:n,\:{x}_{0}$ being the initial state.

Initial Distribution: Suppose $\:\left\{{\text{X}}_{\text{n},}\text{n}\ge\:0\right\}$ is a Markov chain with state space S and,

$$\:{\text{P}}_{\text{i}}^{\left(0\right)}=\text{P}\left[{\text{X}}_{0}=\text{i}\right]\:\text{f}\text{o}\text{r}\:\text{i}\:\text{ϵ}\:\text{S}\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\left(3\right)$$

Then $\:{\text{P}}^{\left(0\right)}=\left\{{\text{P}}_{\text{i}}^{\left(0\right)},\:\text{i}\:\text{ϵ}\:\text{S}\right\}$ is the probability mass function of $\:{\text{X}}_{0}$ and is known as the initial distribution of the Markov chain.

$$\:\text{P}\left[{\text{X}}_{\text{n}+1}=\left.\text{j}\right|{\text{X}}_{\text{n}}=\text{i}\right]\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\left(4\right)$$

Of transition in one step, from state $\:i$ at time $\:n$ to state $\:j$ at time $\:n+1$, is known as the one step transition probability. It is denoted by $\:{\text{P}}_{\text{i}\text{j}}^{\text{n},\text{n}+1}$ is a function of $\:i,j\:and\:n$.

The Ordinary Kriging interpolation technique from ArcGIS is adopted in this study for patterns analysis to 5km, which is based on data from about 27 regions.

3.1 Trends in precipitation and overall likelihood

The monthly trends of precipitation; 1950–2020

Significant variations in rainfall patterns over the decades can be seen in the monthly precipitation trends from 1950 to 2020 (see Fig. 2). 1950–1960: Of the rainy months, May (18.4mm) had the least amount of precipitation, while January (68.5mm) and April (61.4mm) had the most. This period was characterised by relatively high precipitation. 1960–1970: Some months saw less rainfall than others, including January (13.8mm) and May (8.1mm). Still, April saw 68.9mm of rain, which was a significant amount of precipitation. 1970–1980: The amount of rain began to stabilise, with notable levels being maintained in January (42.6mm) and April (58.3mm). Monthly precipitation varied less overall than it did ten years ago. 1980–1990: There was a noticeable drop in precipitation, particularly in February (5.8mm) and March (6.8mm). The values for April and November stayed comparatively stable at 32.9mm and 34.4mm, respectively. 1990–2000: The downward trend in precipitation persisted, with notably low amounts in January (3.6mm) and November (-9.2mm). This time frame points to a concerning pattern of declining precipitation. The years 2000–2010 saw some of the lowest precipitation trends; some months, including January (-1.4mm) and March (-10.1mm), had negative readings. Monthly rainfall was seriously declining, as seen by the overall trend. 2010–2020: Based on result, there appears to have been a minor improvement in precipitation levels over the preceding ten years, especially in April (25.8mm), October (15.4mm), November (15.7mm), and December (13.9mm).

[FIG.2]

Overall chance of rain

The overall chance of rain in March showed variability from 1981 to 2021 (see Fig. 3). In 1981, the chance of rain was 40% by 1991, this decreased slightly to 36%. The percentage remained stable at 36% in 2001. In 2011, it saw a slight increase to 38%. By 2021, the chance of rain returned to 40%. Fluctuations around 36–40%, indicating some stability in precipitation likelihood over the decades, but with no clear long-term trend. The chance of rain in October consistently remained below 22% from 1981 to 2021. Indicates a persistent low likelihood of precipitation during this month. Similar to October, the chances of rain in November also remained below 22% throughout the analysed period from 1981 to 2021. Points to a consistent pattern of low precipitation likelihood in November during the extended November to April period.

[FIG.3]

3.2 Variability and shifts in the dates when rainfall begins

The mean spatial patterns of rainfall onset dates

The mean spatial patterns of rainfall onset dates across various regions of Tanzania during three key months: March (for the March-April-May, or MAM season) (see Fig. 4 (top-left)), October (for the October-November-December, or OND season) (see Fig. 4 (top-right), and November (for the long season, November to April) (see Fig. 4 (bottom)). The onset of rainfall varies regionally, with key findings as follows: Regions with earlier onset (around 63 days): Geita, Kagera, Mwanza, Shinyanga, Simiyu. Moderate onset (65 to 67 days): Morogoro Kaskazini (65), Kilimanjaro (66), Manyara (66), Mara (64), Tanga (69), Dar es salaam (67), Arusha (67). Later onset (above 67 days): Kusini Pemba (69), Kusini Unguja (68). The data indicate that most regions experience rainfall onset at the beginning of March, with coastal regions and islands having a slightly delayed onset compared to the mainland. The rainfall onset during October shows a wider range of days, with the following observations: Ealy onset (around 276–279 days): Kagera (276), Geita (278), Mwanza (279). Later onset (above 280 days): Shinyanga (287), Simiyu (282), Dar es Salaam (282), Kilimanjaro (280), Manyara (289), Mara (288), Morogoro Kaskazini (287). Highest onset: occurs in Arusha (297) and the coastal regions of Kusini Pemba and Kusini Unguja (both at 296). This indicates that central regions may experience slightly delayed rainfall onset in comparison to the coastal and northern regions. Rainfall onset continues into November as part of the long season: Regions such as Lindi (319), Ruvuma (318), and Dodoma (324) experience the latest onset dates. Other notable regions with late onset include Kigoma (307), Tabora (312), and Mbeya (314). This indicates a consistent pattern wherein southern and central regions tend to witness prolonged waiting times for the start of the rainy season.

[FIG.4]

The trend of rainfall onset from the climatology during 1981 to 2023

The rainfall onset dates during the short rain season (March-April-May) from 1981 to 2023 across various regions (see Table 2), reveals several significant findings based on the Mann-Kendall test and the Sens slope estimations. Most regions showed a negative trend in the rainfall onset dates, indicating a tendency for these dates to occur earlier over the observed period. Geita: A significant downward trend with a Mann-Kendall statistic of -249 and a slope of -0.042 days per year, suggesting earlier rainfall onset. Kagera: No significant trend (S= -124; p = 0.176; $\:{\text{Q}}_{2}$=0), indicating stability in onset dates. Dar es Salaam: A significant negative trend (S= -232; p=0.014; $\:{\text{Q}}_{2}$ = -0.098 days per year), further indicating earlier rainfall. Manyara: Showed a very clear significant trend (S=-298; p = 0.002; $\:{\text{Q}}_{2}$=-0.176 days per year) indicating an earlier onset of rain. Several regions such as Shinyanga, Simiyu, Mwanza, and Morogoro exhibited no statistically significant trends (high p-values), indicating that rainfall onset dates in these areas have remained relatively stable over the years. Regions like Kilimanjaro and Mara exhibited trends suggestive of a potential shift but lacked statistical significance (p-values above 0.05).

Table 1 information about the 27 selected representative regions (Meteorological stations) across Tanzania.
No.	Station Name	Latitude (°)	Longitude (°)	Altitude (m)
1	Geita	-2.8	32.2	1240
2	Iringa	-7.7	35.7	1428
3	Kagera	-1.3	31.8	1144
4	Katavi	-6.3	31.2	1502
5	Mwanza	-2.5	32.9	1140
6	Njombe	-9.3	34.8	1821
7	Rukwa	-8.0	31.4	800
8	Shinyanga	-3.6	33.4	1127
9	Simiyu	-2.8	34.1	1377
10	Arusha	-3.4	36.6	1387
11	Dar-es-salaam	-6.8	39.2	53
12	Dodoma	-6.2	35.8	1120
13	Kigoma	-4.9	29.6	820
14	Kilimanjaro	-3.0	37.3	813
15	Lindi	-10.0	39.7	402
16	Manyara	-4.3	36.9	1197
17	Mara	-1.5	33.8	1147
18	Mbeya	-8.9	33.4	1758
19	Morogoro	-6.8	37.6	526
20	Mtwara	-10.2	40.1	113
21	Tanga	-5.0	39.0	49
22	Ruvuma	-10.6	36.2	1036
23	Singida	-4.8	34.7	1260
24	Kusini Unguja	-6.0	39.2	27
25	Kusini Pemba	-5.1	39.4	4
26	Tabora	5.1	32.4	1200
27	Mahenge	-8.4	36.4	1400

Table 2 shows the trend of the rainfall onset dates throughout the rainy season (March–April–May (MAM), October–November–December (OND) and November to April (N-A)) as determined by Mann-Kendall's test and Sen's slope (𝒬2)
Region	Variable	MAM	OND	N-A
Geita	S	-249	-173	-
	p	0.007	0.067	-
	𝒬₂	-0.042	-0.045	-
Kagera	S	-124	-110	-
	p	0.176	0.240	-
	𝒬₂	0	0	-
Mwanza	S	-176	-53	-
	p	0.060	0.582	-
	𝒬₂	-0.04	0	-
Shinyanga	S	-54	-57	-
	p	0.569	0.557	-
	𝒬₂	0	-0.064	-
Simiyu	S	41	-54	-
	p	0.669	0.578	-
	𝒬₂	0	-0.033	-
Arusha	S	-145	-188	-
	p	0.13	0.050	-
	𝒬₂	-0.094	-0.333	-
Dar-es salaam	S	-232	-36	-
	p	0.014	0.714	-
	𝒬₂	-0.098	-0.053	-
Kilimanjaro	S	-76	-54	-
	p	0.429	0.575	-
	𝒬₂	-0.026	0	-
Manyara	S	-298	-98	-
	p	0.002	0.308	-
	𝒬₂	-0.176	-0.095	-
Mara	S	-133	-176	-
	p	0.161	0.066	-
	𝒬₂	-0.027	-0.273	-
Morogoro	S	-52	-52	-
	p	0.593	0.593	-
	𝒬₂	-0.04	-0.04	-
Kusini Pemba	S	-127	-285	-
	p	0.185	0.003	-
	𝒬₂	-0.111	-0.545	-
Kusini Unguja	S	-179	69	-
	p	0.061	0.476	-
	𝒬₂	-0.154	0.072	-
Tanga	S	-78	-6	-
	p	0.419	0.958	-
	𝒬₂	-0.074	0	-

Table 3 shows the trend of the rainfall onset dates throughout the rainy season (March–April–May (MAM), October–November–December (OND) and November to April (N-A)) as determined by Mann-Kendall's test and Sen's slope (𝒬2)
Region	Variable	MAM	OND	N-A
Lindi	S	-	-	39
	p	-	-	0.690
	𝒬₂	-	-	0.037
Ruvuma	S	-	-	30
	p	-	-	0.761
	𝒬₂	-	-	0.031
Mahenge	S	-	-	-68
	p	-	-	0.482
	𝒬₂	-	-	-0.059
Kigoma	S	-	-	25
	p	-	-	0.792
	𝒬₂	-	-	0
Tabora	S	-	-	-84
	p	-	-	0.382
	𝒬₂	-	-	-0.056
Dodoma	S	-	-	-123
	p	-	-	0.201
	𝒬₂	-	-	-0.143
Singida	S	-	-	-155
	p	-	-	0.1059
	𝒬₂	-	-	-0.167
Iringa	S	-	-	-20
	p	-	-	0.842
	𝒬₂	-	-	0
Mbeya	S	-	-	-15
	p	-	-	0.883
	𝒬₂	-	-	0

Rainfall onset dates during the short rain season (October-November-December, OND) (see Table 2), has revealed significant spatial variations across different regions. Kusini Pemba, showed a significant downward trend (S = -285) with a low significance level (p = 0.003), indicating an early onset of the rains. Other regions, such as Kagera, Mara, and Arusha, exhibited statistically significant trends with varying degrees of early onset, but these trends were weaker compared to Kusini Pemba. In contrast, regions like Tanga indicated no significant change in onset dates (p = 0.958), suggesting stability in rainfall patterns. The Sens slope values ($\:{\text{Q}}_{2}$) indicate the rate of change in rainfall onset. For instance, Kusini Pemba had a steep slope of -0.545 days per year, indicating a substantial shift in rainfall occurrence. This contrasts with regions like Simiyu and Mwanza, which also demonstrated earlier onset but with modest slopes, reflective of slight changes. The study underscores the heterogeneity of rainfall patterns across the surveyed regions. Some areas, like Dar es Salaam, demonstrated a high degree of variability in onset dates (S = -36, p = 0.714), but still experienced shifts earlier than historical norms.

[Table 2]

The analysis presented in Table 3 examines the trend of rainfall onset dates during the long rain season (November-April) from 1981 to 2023. Lindi: A slight delay in rainfall onset is observed (39 days), with a strong positive trend (p = 0.690) indicating that the onset dates are becoming earlier over time at a rate of 0.037 days per year. Ruvuma: Similar to Lindi, Ruvuma also shows a positive trend with a delay of 30 days (p = 0.761) and a modest rate of change (0.031 days per year). In contrast, Mahenge shows an insignificant delay in onset by 68 days (p = 0.482), with a slight decrease in onset dates (-0.059 days per year). Kigoma: The onset date is advanced not significant by 25 days, with a p-value 0.792. Tabora exhibits an insignificant delay of 84 days (p = 0.382) and a negative trend (-0.056 days per year). Dodoma: A larger delay of 123 days with a very weak positive trend (p = 0.201) suggests minimal change in onset dates, showing a pronounced negative slope (-0.143 days per year). Notably, Singida has the most considerable delay of 155 days with an insignificant (p = 0.1059) and negative trend (-0.167 days per year). The rainfall onset date in Iringa is relatively stable, with a slight delay of 20 days and a strong positive trend (p = 0.842). Similar to Iringa, Mbeya also shows minimal delay (15 days) and an insignificant positive trend (p = 0.883). Generally, the findings indicate variability in the trends of rainfall onset across different regions, with some areas experiencing delays and others showing a trend towards earlier onset dates. The trends are characterized by a mix of significant positive and negative changes, suggesting a complex response to climatic influences over the studied period.

[Table 3]

Fitting a Markov chain model to data

The probability of rain for various regions during the March-April-May (MAM) period (refer to Fig. 5), reveal notable variations based on previous weather conditions. The analysis splits the probabilities into two categories: the likelihood of rain given that the previous day was dry (rd), and the likelihood of rain given that the previous day was rainy (rr). March findings: Regions like Simiyu (0.64) and Arusha (0.57) show relatively high probabilities of rain when rain occurred the previous day. The lowest probabilities of rain following a dry day are found in Tanga (0.11) and Manyara (0.13), indicating a lower likelihood of precipitation in these areas at the start of the MAM season. Kagera has the highest chance of rain following a dry day (0.46), while Dar es salaam has the lowest (0.18). April findings: The probabilities generally increase compared to March. For example, Kagera's (rr) increases to 0.67, indicating a significant likelihood of continued rainfall after a rainy day. Regions like Mara (0.72) and Geita (0.58) show strong (rr) probabilities, suggesting that once it rains, it’s more likely to rain again the following day. The chances of rain after a dry day (rd) are higher across the board compared to March, particularly in Kagera (0.54) and Geita (0.39). May findings: There is a noticeable decrease in probabilities for both (rd) and (rr), with regions like Shinyanga (0.10) showing particularly low potential for rain after a dry day. The likelihood of rain following a rainy day remains relatively stable, with Kagera (0.66) and Mara (0.66) still demonstrating significant rain continuation patterns. Compared with April, many regions show lower (rd) probabilities, such as Mwanza (0.24) and Simiyu (0.24), suggesting a drying trend towards the end of the MAM seasons.

October marks the beginning of the short rainy season (OND) (see Fig. 5) in many regions in East Africa. The observed probabilities suggest a transitional phase, where the likelihood of rain varies significantly by region. For example, Geita (0.50) and Kagera (0.47) have notably high probabilities. Shinyanga (0.04) and Manyara (0.04) exhibit very low chances, indicating less likelihood of rain after a dry day. Most regions show an increase in the probability of rain when compared to October. Kagera (0.54) and Simiyu (0.58) show the strongest likelihood of rain following a rainy day. Manyara (0.09) and Morogoro (0.09) continues to demonstrate minimal chances. Many regions maintain or see a rise in probabilities, especially rainy days. Kagera (0.60) and Simiyu (0.61) indicate the highest probabilities after a rainy day, suggesting a strong trend of continuation in precipitation. Regions like Manyara (0.12) and Dar es Salaam (0.13) remain on the lower end for dry day probabilities.

[FIG.5]

Furthermore, the results (see Fig. 6) present a clear trend of increasing probabilities of rain across various regions from November to February during the long rainy season. This pattern reflects seasonal climatic behaviour typical in tropical regions, where initial monthly rainfall is generally lower before escalating as the wet season intensifies. Kigoma and Ruvuma: Exhibit the highest probabilities throughout the analysis period. For instance, Kigoma's probability after a rain day (rr) peaked at 0.63 in December and remained high throughout January and February, indicating a reliable likelihood of consecutive rainy days. This consistency suggests that these regions may experience more stable precipitation patterns, making them critical for agricultural planning and water resource management. Dodoma: Stands out with the lowest values for the probability of rain, both after dry days and rainy days. Its (rd) values, particularly in November (0.03), suggest that Dodoma is less likely to receive rain after a dry spell, highlighting its drier climate. The significant increases from November through February highlight the seasonal shift towards the peak rainy period, often observed in equatorial climates. December often marks a transitional point with uplifted probabilities, moving towards the heights observed in February. Specifically, February shows the highest values in both rainy (rr) and dry day (rd) conditions across many regions, with Ruvuma reaching 0.77 after a rain day. This elevates concerns about flooding and waterlogging, especially in regions with higher probability rates.

March: The probability of rain given that the previous day was dry (rd) varies across regions, with Kigoma showing the highest probability at 0.53 and Dodoma the lowest at 0.23. When the previous day was rainy (rr), Ruvuma exhibits the highest probability at 0.72, while Lindi stands at 0.61. Most regions demonstrate a significant likelihood of continued rainfall following a rainy day. The generally higher probabilities for (rr) compared to (rd) suggest a strong tendency for rainy days to cluster together. April: The probabilities of rain after a dry day (rd) decrease across all regions compared to March. Kigoma still has the highest at 0.56, but Dodoma's probability drops further to 0.13. The probabilities after a rainy day (rr) also decrease compared to March, with Ruvuma at 0.48 being the lowest recorded. Overall, the month of April tends to show lower probabilities for both scenarios, indicating a possible transition from the rainy season towards drier conditions.

[FIG.6]

Global warming leads to shifts in weather patterns, including changes in precipitation (Dore 2005; Rahat et al. 2024; Tabari 2020). Increased temperatures can alter the hydrological cycle, resulting in more intense rainfall in certain months (Rahat et al. 2024). El Niño and La Niña Events, these climate phenomena influence weather patterns worldwide. El Niño can cause heavier rainfall, while La Niña may lead to drier conditions (Cai et al. 2014; Hund et al. 2021; Luo et al. 2010). Variations in these cycles could impact the observed rain chances in Tanzania (Mbigi and Xiao 2021). Deforestation, urbanization, and agricultural expansion can affect local climates, altering rainfall distribution and intensity (Ellison et al. 2017). Changes to the landscape may influence evaporation rates and cloud formation (Liu et al. 2020). The shifting patterns in ocean currents and temperatures around the Indian Ocean can directly affect rainfall patterns over East Africa, including Tanzania (Endris et al. 2019; Khan et al. 2021; Xiao et al. 2022). Changes in the jet stream or monsoon systems can lead to variations in rainfall (Liu et al. 2019; Pratte et al. 2024). Shifts in wind patterns may impact how moisture is transported to Tanzania. Natural variability in weather patterns can cause fluctuations in rainfall amounts from year to year, affecting long-term averages and trends.

Areas like Kilimanjaro, Arusha, and Manyara (Northern Tanzania) experience significant variations in rainfall, especially during the long rainy season. Changes affect agriculture and water supply. Regions such as Kigoma and Tabora (Western and central western Tanzania) face alterations in rainfall patterns, influencing crop yields and food security in these agricultural zones. Southern Highlands: This area, including Mbeya and Rukwa, is also experiencing shifts in precipitation patterns, affecting the economy reliant on agriculture and livestock. Places like Dar es Salaam and Tanga (Coastal regions) are affected by rising sea levels and changing rainfall, which can lead to flooding and impact freshwater resources. Regions such as Dodoma and Morogoro (Eastern and Central Tanzania) see fluctuations in seasonal rainfall that can disrupt traditional farming practices. Areas around Lake Victoria, including Mwanza, are facing impacts from changing rainfall patterns, affecting fishery resources and agricultural activities.

The findings from the research indicate a mixed picture regarding precipitation trends over the specified years. March shows some variability, with a return to a higher likelihood of rain (40%) by 2021 after a decade of lower percentages. This could suggest resilience in certain years against drier conditions or shifts in seasonal weather patterns influencing March rainfall. In contrast, both October and November display a consistent trend of low precipitation likelihood, suggesting that these months are generally dry and less prone to significant rainfall. This could be critical for water management, agriculture, and planning for these seasons since low chances of rain may result in a reliance on stored water or other sources during these months.

The spatial variability in rainfall onset dates highlights the complexity of Tanzania's climatology. Understanding these patterns is crucial for agriculture, water resource management, and disaster preparedness, as the timing and amount of rainfall significantly influence crop yields and water availability. Regional differences in onset dates suggest that local climatic factors, such as elevation, proximity to water bodies, and prevailing winds, play significant roles in determining rainfall patterns. The coastal regions tend to receive rainfall slightly earlier, possibly due to moisture-laden winds from the ocean, while inland areas have more variable onset times. Significant trends in earlier rainfall onset dates were found in Geita, Dar es Salaam, and Manyara, indicating a possible shift in climatic patterns that may affect agricultural practices and water resource management in these regions. Regions like Kagera, Shinyanga, and Simiyu showed no significant trends, suggesting regional variability in rainfall patterns that may influence local ecosystems differently. An earlier onset of rains could lead to changes in crop planting schedules, which could be beneficial for some crops but detrimental for others, possibly leading to reduced yields. Farmers in affected areas will need to adopt their practices accordingly. The observed trends may correlate with climate change effects, necessitating further research to understand the underlying causes. The variability and significance of trends across regions emphasize the need for localized climate adaptation strategies.

Furthermore, the findings highlight significant variations in rainfall behaviours based on preceding weather conditions. The transition from March to May shows a notable increase in rainfall probabilities in April, many of which align with established climatological trends where transition months (like April) often serve as critical periods for seasonal rains (Nicholson 2017). The continuation of rainfall after previous rain days (high rr values) suggests that once precipitation begins, the moisture often leads to a favourable feedback loop that sustains further rain (Bates et al. 2013). Studies have shown that such persistence of rainy days can be linked to the regional moisture availability, local geography (e.g., altitude and proximity to bodies of water), and broader climatic systems, such as the Inter-Tropical Convergence Zone (ITCZ) (Chao 2001; Kawase et al. 2019; Koster et al. 2003). Regions like Kagera and Mara exhibit strong(rr) values, indicating that local climates support sustained moisture, which can be crucial for agriculture and hydrology. The stark differences in rain probability, particularly following dry days, reflect underlying climatic variability and can have direct implications for farming practices and water resource management. For instance, Tanga's low temperatures (both rd and rr) indicate a potential desertification trend or insufficient rains, emphasizing the need for adaptive agricultural strategies (Colloff et al. 2021). Conversely, regions with consistently high rainfall probabilities could benefit from advanced irrigation techniques and sustainable water management practices to optimize agricultural output.

The probability of rain following dry days (rd) is notably low in regions like Shinyanga and Manyara, indicating that rainfall is more likely to be generated from existing moisture in the atmosphere rather than new systems forming after a dry spell. This aligns with meteorological observations that rain often follows conditions of humidity and can lead to patterns of localized weather phenomena. From October to December, there is a noticeable increase in probabilities of rain across most regions following both dry and rainy days. This pattern suggests that as the rainy season progresses, the atmosphere becomes more conducive to precipitation, reinforcing the role of moisture accumulation from previous storms. Such trends are vital for anticipating agricultural cycles and planning for food security. The diverse rainfall probabilities reflect geographical and ecological differences among the regions. For instance, coastal areas like Tanga may experience different climatic influences due to proximity to the ocean compared to inland regions like Mwanza or Shinyanga, which may be more susceptible to continental climate effects.

Understanding rainfall probabilities aids farmers in timing planting and harvesting, which is critical for maximizing yields. Regions with high probabilities of rain after rainy days are likely to be more reliable for crops that depend on consistent moisture. Conversely, areas with low probabilities may necessitate strategies like irrigation. With rising probabilities, especially in areas like Manyara, Arusha, and Singida water resource management strategies should be reviewed and optimized to reinforce resilience against potential drought conditions in dry months. Monitoring rainfall trends can inform resource allocation and conservation practices, particularly for vulnerable communities.

Given that the rainy season tends to peak in February, farmers can optimize crop planting and harvesting schedules based on this data. Regions with consistent rainfall patterns like Kigoma and Ruvuma can focus on crops that thrive during the wet season, whereas areas like Dodoma may need drought-resistant strategies. The variability in rainfall probabilities underscores the need for effective water resource management. High probabilities in certain periods indicate opportunities for rainwater harvesting and storage systems to combat periods of drought seen in drier regions. With ongoing climate change, understanding these traditional weather patterns is essential for adaptation strategies. Increased extremes in rainy and dry spells could shift probabilities, necessitating continuous monitoring and responsive adaptation measures. Regions with high probabilities of consecutive rainy days need to prepare for associated risks such as landslides, flooding, and associated impacts on infrastructure and human health.

According to the results, March precipitation shows moderate variability with a recent return to greater probability, while October and November continue to show low rainfall probabilities though there appears to have been a minor improvement. These tendencies underscore the need for more research on long-term climate trends and their effects, as they may have important ramifications for the region's agricultural planning, water resource management, and climate adaptation efforts. The results highlight the significance of regional weather forecasting and adaptation plans for Tanzanian agriculture and water resources. Comprehending the average spatial patterns of rainfall onset might help stakeholders and policymakers create efficient strategies to lessen the effects of erratic rainfall. Future studies should investigate the underlying climate conditions driving these patterns in greater detail.

By analysing rainfall probabilities for the months of March, April, and May, this study reveals important trends in the climate and their implications for areas that are vulnerable. According to the results, April is a critical month for precipitation, with some areas exhibiting a noticeably higher chance of precipitation after prior rainy days. However, many regions show decreased rainfall possibilities as May draws near, calling for proactive and flexible methods to water management and agriculture. Given the anticipated effects of climate change on precipitation variability and agricultural productivity worldwide, efficient planning based on these patterns can significantly improve crop yields and conserve water resources. In order to promote resilience in impacted areas and guarantee sustainable resource management going forward, stakeholders should make use of these insights. But, from October to December in various regions has highlighted significant trends that can be crucial for food security, agricultural planning, and resource management. Regions like Kagera and Simiyu show strong prospects for rain after previous rainy days, while areas like Manyara and Shinyanga present ongoing challenges with low probabilities after dry days.

The analysis highlights significant variations in the probability of rainfall across Tanzania during the long rainy season from November to April. The observed increase in rainfall chances as the season progresses underlines the importance of adaptive strategies for agriculture, water resource management, and disaster mitigation. Understanding and analysing weather trends, such as those observed in this study, is instrumental in preparing for the impacts of climate variability. Stakeholders, including farmers, local governments, and policymakers, should leverage this data to make informed decisions that promote resilience and sustainability in the face of changing climatic conditions. Setting up monitoring systems and predictive models can enhance readiness for both expected rain events and potential droughts. Ultimately, these measures are foundational to ensuring food security, economic stability, and environmental sustainability in the affected regions.

Acknowledgments: The first author would like to appreciate Prof. Yunsheng Lou for his trust in my ideas and abilities and his supervision.

Conflict of Interest statement: The authors declare that have no conflict of interest

Funding statement: The authors have no relevant financial or non-financial interest to disclose

Authors’ contributions: The data search and Formal analysis, Methodology, Software, Visualization, Writing - original draft, Writing - review & editing by Dawido Simtenda Magang and supervision by Prof. Yunsheng Lou.

Data availability statement: The data generated during analysis are available via request from email:

[email protected] and https://data.chc.ucsb.edu/products/CHIRPS-2.0/ website.

Arndt C, Farmer W, Strzepek K, and Thurlow J (2012) Climate Change, Agriculture and Food Security in Tanzania. Review of Development Economics, 378–393. https://doi.org/10.1111/j.1467-9361.2012.00669.x
Basalirwa C P K, Odiyo J O, Mngodo R J and Mpeta E J (1999) The climatological regions of Tanzania based on the rainfall characteristics. International Journal of Climatology, 69–80. https://doi.org/10.1002/(sici)1097-0088(199901)19:1<69::aid-joc343>3.0.co;2-m
Bates B C, Kundzewicz, Zbigniew, Wu Shaohong, and Palutikof Jean (2013) Climate change and water. Intergovernmental Panel on Climate Change.
Byrne M P, Pendergrass A G, Rapp A D, and Wodzicki K R (2018) Response of the Intertropical Convergence Zone to Climate Change: Location, Width, and Strength. In Current Climate Change Reports (pp. 355–370). Springer. https://doi.org/10.1007/s40641-018-0110-5
Cai W, Borlace S, Lengaigne M, Van Rensch P, Collins M, Vecchi G, Timmermann A, Santoso A, Mcphaden M J, Wu L, England M H, Wang G, Guilyardi E, and Jin F F (2014) Increasing frequency of extreme El Niño events due to greenhouse warming. Nature Climate Change, 111–116. https://doi.org/10.1038/nclimate2100
Chao W C (2001) Multiple Quasi Equilibria of the ITCZ and the Origin of Monsoon Onset. Part II: Rotational ITCZ Attractors.
Colloff M J, Gorddard R, Abel N, Locatelli B, Wyborn C, Butler J R A, Lavorel S, van Kerkhoff L, Meharg S, Múnera-Roldán C, Bruley E, Fedele G, Wise R M, and Dunlop M (2021) Adapting transformation and transforming adaptation to climate change using a pathways approach. Environmental Science and Policy, 163–174. https://doi.org/10.1016/j.envsci.2021.06.014
da Silva R M, Santos C A G, Moreira M, Corte-Real J, Silva V C L, and Medeiros I C (2015) Rainfall and river flow trends using Mann–Kendall and Sen’s slope estimator statistical tests in the Cobres River basin. Natural Hazards, 1205–1221. https://doi.org/10.1007/s11069-015-1644-7
Dieppois B, Capotondi A, Pohl B, Chun K P, Monerie P A, and Eden J (2021) ENSO diversity shows robust decadal variations that must be captured for accurate future projections. Communications Earth and Environment. https://doi.org/10.1038/s43247-021-00285-6
Dinku T, Funk C, Peterson P, Maidment R, Tadesse T, Gadain H, and Ceccato P (2018) Validation of the CHIRPS satellite rainfall estimates over eastern Africa. Quarterly Journal of the Royal Meteorological Society, 292–312. https://doi.org/10.1002/qj.3244
Dore M H I (2005) Climate change and changes in global precipitation patterns: What do we know? In Environment International (pp. 1167–1181). https://doi.org/10.1016/j.envint.2005.03.004
Ellison D, Morris C E, Locatelli B, Sheil D, Cohen J, Murdiyarso D, Gutierrez V, Noordwijk M van, Creed I F, Pokorny J, Gaveau D, Spracklen D V, Tobella A B, Ilstedt U, Teuling A J, Gebrehiwot S G, Sands D C, Muys B, Verbist B, … Sullivan C A (2017) Trees, forests and water: Cool insights for a hot world. Global Environmental Change, 51–61. https://doi.org/10.1016/j.gloenvcha.2017.01.002
Endris H S, Lennard C, Hewitson B, Dosio A, Nikulin G, and Artan G A (2019) Future changes in rainfall associated with ENSO, IOD and changes in the mean state over Eastern Africa. Climate Dynamics, 2029–2053. https://doi.org/10.1007/s00382-018-4239-7
Feller W (1959) An introduction to Probability Theory and Applications.
Fowler H J, and Wilby R L (2010) Detecting changes in seasonal precipitation extremes using regional climate model projections: Implications for managing fluvial flood risk. Water Resources Research. https://doi.org/10.1029/2008WR007636
Funk C, Peterson P, Landsfeld M, Pedreros D, Verdin J, Shukla S, Husak G, Rowland J, Harrison L, Hoell A, and Michaelsen J (2015) The climate hazards infrared precipitation with stations - A new environmental record for monitoring extremes. Scientific Data. https://doi.org/10.1038/sdata.2015.66
Hamed K H (2008) Trend detection in hydrologic data: The Mann-Kendall trend test under the scaling hypothesis. Journal of Hydrology, 350–363. https://doi.org/10.1016/j.jhydrol.2007.11.009
Hamed K H, and Rao A R (1998) Hydrology A modified Mann-Kendall trend test for autocorrelated data. In Journal of Hydrology.
Harrison L, Landsfeld M, Husak G, Davenport F, Shukla S, Turner W, Peterson P, and Funk C (2022) Advancing early warning capabilities with CHIRPS-compatible NCEP GEFS precipitation forecasts. Scientific Data. https://doi.org/10.1038/s41597-022-01468-2
Hu Z, Liu S, Zhong G, Lin H, and Zhou Z (2020) Modified Mann-Kendall trend test for hydrological time series under the scaling hypothesis and its application. Hydrological Sciences Journal, 2419–2438. https://doi.org/10.1080/02626667.2020.1810253
Hund S V, Grossmann I, Steyn D G, Allen D M, and Johnson M S (2021) Changing Water Resources Under El Niño, Climate Change, and Growing Water Demands in Seasonally Dry Tropical Watersheds. Water Resources Research. https://doi.org/10.1029/2020WR028535
Indeje M, Semazzi F H M, and Ogallo L J (2000) ENSO signals in East African rainfall seasons. International Journal of Climatology, 19–46. https://doi.org/10.1002/(SICI)1097-0088(200001)20:1<19::AID-JOC449>3.0.CO;2-0
Jiqin H, Gelata F T, and Gemeda S C (2023) Application of MK trend and test of Sen’s slope estimator to measure impact of climate change on the adoption of conservation agriculture in Ethiopia. Journal of Water and Climate Change, 977–988. https://doi.org/10.2166/wcc.2023.508
Jolliffe I T, and Sarria‐dodd D E (1994) Early detection of the start of the wet season in tropical climates. International Journal of Climatology, 71–76. https://doi.org/10.1002/joc.3370140106
Jones K, Nowak A, Berglund E, Grinnell W, Temu E, Paul B, Renwick L L R, Steward P, Rosenstock T S, and Kimaro A A (2023) Evidence supports the potential for climate-smart agriculture in Tanzania. Global Food Security. https://doi.org/10.1016/j.gfs.2022.100666
Jury M R (2024) Climate variability and hydrology impacts in east Africa’s Rwenzori Mountains. Journal of Hydrology: Regional Studies. https://doi.org/10.1016/j.ejrh.2024.101922
Kanga S, Suraj, Singh K, Shevkani K, Pathak V, and Sajan B (2024) World Sustainability Series Transforming Agricultural Management for a Sustainable Future Climate Change and Machine Learning Perspectives.
Kavishe G M, Tilwebwa P, and Limbu S (2020) Variation of October to December rainfall in Tanzania and its association with sea surface temperature. https://doi.org/10.1007/s12517-020-05535-z/Published
Kawase H, Imada Y, Sasaki H, Nakaegawa T, Murata A, Nosaka M, and Takayabu I (2019) Contribution of Historical Global Warming to Local-Scale Heavy Precipitation in Western Japan Estimated by Large Ensemble High-Resolution Simulations. Journal of Geophysical Research: 6093–6103. https://doi.org/10.1029/2018JD030155
Kebacho L L (2021) Anomalous circulation patterns associated with 2011 heavy rainfall over northern Tanzania. Natural Hazards, 2295–2312. https://doi.org/10.1007/s11069-021-04920-5
Khan S, Piao S, Zheng G, Khan I U, Bradley D, Khan S, and Song Y (2021) Sea surface temperature variability over the tropical Indian Ocean during the ENSO and IOD Events in 2016 and 2017. Atmosphere. https://doi.org/10.3390/atmos12050587
Koster R D, Chang Y, Wang H, and Schubert S D (2003) Impacts of Local Soil Moisture Anomalies on the Atmospheric Circulation Global Modeling and Assimilation Office. In A Comprehensive Analysis over North America.
Kukal M S, and Irmak S (2018) Climate-Driven Crop Yield and Yield Variability and Climate Change Impacts on the U.S. Great Plains Agricultural Production. Scientific Reports. https://doi.org/10.1038/s41598-018-21848-2
Liu Z, Cheng L, Zhou G, Chen X, Lin K, Zhang W, Chen X, and Zhou P (2020) Global Response of Evapotranspiration Ratio to Climate Conditions and Watershed Characteristics in a Changing Environment. Journal of Geophysical Research. https://doi.org/10.1029/2020JD032371
Liu Z, Ming Y, Wang L, Bollasina M, Luo M, Lau N C, and Yim S H L (2019) A Model Investigation of Aerosol-Induced Changes in the East Asian Winter Monsoon. Geophysical Research Letters, 10186–10195. https://doi.org/10.1029/2019GL084228
Luo J J, Zhang R, Behera S K, Masumoto Y, Jin F F, Lukas R, and Yamagata T (2010) Interaction between El Niño and extreme Indian Ocean dipole. Journal of Climate, 726–742. https://doi.org/10.1175/2009JCLI3104.1
MacLeod D (2018) Seasonal predictability of onset and cessation of the east African rains. Weather and Climate Extremes,27–35. https://doi.org/10.1016/j.wace.2018.05.003
Mann H B (1945) Nonparametric Tests Against Trend.
Mapande A T, and Reason C J C (2005) Links between rainfall variability on intraseasonal and interannual scales over western Tanzania and regional circulation and SST patterns. Meteorology and Atmospheric Physics, 215–234. https://doi.org/10.1007/s00703-005-0130-2
Mason S J, and Goddard L (2001) Probabilistic precipitation anomalies associated with ENSO. Bulletin of the American Meteorological Society, 619–638. https://doi.org/10.1175/1520-0477(2001)082<0619:PPAAWE>2.3.CO;2
IPCC (2021) Working Group I Contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. www.ipcc.ch
Mbigi D, Onyango A O, Mtewele Z F, Kiprotich P, and Xiao Z (2022) Coupled Model Intercomparison Project Phase 6 simulations of the spatial structure of rainfall variability over East Africa: Evaluation and projection. International Journal of Climatology, 9865–9885. https://doi.org/10.1002/joc.7868
Mbigi D, and Xiao Z (2021) Tanzanian rainfall responses to El Niño and positive Indian Ocean Dipole events during 1951–2015. Atmospheric and Oceanic Science Letters. https://doi.org/10.1016/j.aosl.2021.100093
McSweeney M (2009) UNDP Climate Change Contry Profiles Tanzania. Retrieved from https://www.tnrf.org/files/E-INFO-UNDP_Climate_Change_Tanzania_Profile.pdf
Nicholson S E (2017) Climate and climatic variability of rainfall over eastern Africa. Reviews of Geophysics, 590–635. https://doi.org/10.1002/2016RG000544
Odenkunle T O (2004) Rainfall and the length of the growing season in Nigeria. International Journal of Climatology, 467–479. https://doi.org/10.1002/joc.1012
Palmer P I, Wainwright C M, Dong B, Maidment R I, Wheeler K G, Gedney N, Hickman J E, Madani N, Folwell S S, Abdo G, Allan R P, Black E C L, Feng L, Gudoshava M, Haines K, Huntingford C, Kilavi M, Lunt M F, Shaaban A, and Turner A G (2023) Drivers and impacts of Eastern African rainfall variability. In Nature Reviews Earth and Environment (pp. 254–270). Springer Nature. https://doi.org/10.1038/s43017-023-00397-x
Pratte S, Bao K, Li C, Zhang W, Le Roux G, Li G, and De Vleeschouwer F (2024) East Asian monsoon and westerly jet driven changes in climate and surface conditions in the NE drylands of China since the Late Pleistocene. Quaternary Science Reviews, 331. https://doi.org/10.1016/j.quascirev.2024.108637
Rahat S H, Saki S, Khaira U, Biswas N K, Dollan I J, Wasti A, Miura Y, Bhuiyan M A E, and Ray P (2024) Bracing for impact: how shifting precipitation extremes may influence physical climate risks in an uncertain future. Scientific Reports. https://doi.org/10.1038/s41598-024-65618-9
Rowhani P, Lobell D B, Linderman M, and Ramankutty N (2011) Climate variability and crop production in Tanzania. Agricultural and Forest Meteorology, 449–460. https://doi.org/10.1016/j.agrformet.2010.12.002
Segele Z T, and Lamb P J (2005) Characterization and variability of Kiremt rainy season over Ethiopia. Meteorology and Atmospheric Physics, 153–180. https://doi.org/10.1007/s00703-005-0127-x
Shaw A B (1987) An analysis of the rainfall regimes on the coastal region of Guyana. Journal of Climatology, 291–302. https://doi.org/10.1002/joc.3370070307
SOFF Operational Manual Systematic Observations Financing Facility (2021).
TMA Statement on the Status of Tanzania Climate in 2022 (2023).
TMA Statement on the Status of Tanzania Climate in 2023(2024).
Stern R, Knock J, Grayer C, and Leidi S (2002) Introduction to Instat+
Šūmane S, Kunda I, Knickel K, Strauss A, Tisenkopfs T, Rios I des I, Rivera M, Chebach T, and Ashkenazy A (2018) Local and farmers’ knowledge matters! How integrating informal and formal knowledge enhances sustainable and resilient agriculture. Journal of Rural Studies, 232–241. https://doi.org/10.1016/j.jrurstud.2017.01.020
Tabari H (2020) Climate change impact on flood and extreme precipitation increases with water availability. Scientific Reports. https://doi.org/10.1038/s41598-020-70816-2
Tanzania Agriculture Public Expenditure Review (2022).
Technological Approaches for Climate Smart Agriculture (2024) In Technological Approaches for Climate Smart Agriculture. Springer International Publishing. https://doi.org/10.1007/978-3-031-52708-1
With a Special Section on the Role of ICT (2020) http://www.worldbank.org/tanzania/economicupdate
Xiao H M, Lo M H, and Yu J Y (2022) The increased frequency of combined El Niño and positive IOD events since 1965s and its impacts on maritime continent hydroclimates. Scientific Reports. https://doi.org/10.1038/s41598-022-11663-1
Xu H, Chen H, and Wang H (2022) Detectable Human Influence on Changes in Precipitation Extremes Across China. Earth’s Future. https://doi.org/10.1029/2021EF002409
Yang D, Yang X Q, Ye D, Sun X, Fang J, Chu C, Feng T, Jiang Y, Liang J, Ren X, Zhang Y, and Tang Y (2018) On the Relationship Between Probabilistic and Deterministic Skills in Dynamical Seasonal Climate Prediction. Journal of Geophysical Research: Atmospheres, 5261–5283. https://doi.org/10.1029/2017JD028002
Yue S, Pilon P, and Cavadias G (2002) Power of the Mann±Kendall and Spearman’s rho tests for detecting monotonic trends in hydrological series. www.elsevier.com/locate/jhydrol
Yue S, Pilon P, and Phinney B (2003) Canadian streamflow trend detection: Impacts of serial and cross-correlation. Hydrological Sciences Journal, 51–63. https://doi.org/10.1623/hysj.48.1.51.43478

No competing interests reported.

Analysing Changes in Rainfall Dynamics: Onset and Precipitation Probability in Tanzania

Status:

Version 1

Abstract

Figures

1 Introduction

2 Methodology

2.1 Study Area

2.2 Data collection

2.3 Criteria for the assessment of rainfall Onset and Statistical techniques

2.4 Determining Precipitation Probability (Markov chain models)

3 Results

3.1 Trends in precipitation and overall likelihood

3.2 Variability and shifts in the dates when rainfall begins

4 Discussion

5 Conclusion

Declarations

References

Additional Declarations

Status:

Version 1