Long-Term Innovative Trend Analysis of Hydro-Climatic Data of the Sudd Region of South Sudan

Abstract

Floods and droughts are natural disasters that disrupt livelihoods and destroy the environment, with floods constituting up to 40% of all natural disasters globally. South Sudan has experienced severe, recurrent flooding for decades, with two-thirds of the country affected. An integrated flood management system is urgently needed to mitigate impacts and improve community resilience. This requires understanding the inundation process and analyzing flood causes and characteristics. This research leverages data from the Climate Hazards Center InfraRed Precipitation with Station (CHIRPS v2.0) to examine rainfall patterns and analyze trends in annual total precipitation (PRCPTOT), days with precipitation ≥ 20 mm (R20 mm), and simple precipitation intensity (SDII) at the basin scale. It also incorporates Nile River flow data from the Mangala station and Lake Victoria water levels from satellite altimetry. Findings indicate decreasing trends in PRCPTOT, R20 mm, and SDII in Jonglei and Unity States, but increasing trends in river flows and Lake Victoria levels. The Global Surface Water dataset reveals increased water surface areas in these states. These findings suggest that river flow trends oppose rainfall patterns, indicating that local rainfall is not the primary contributor to the recurrent flooding in the area.

1. Introduction

Floods and droughts are among the most disruptive water-related natural disasters, affecting various environmental features and human activities. They compromise wildlife habitats, destroy vegetation, and disrupt agriculture and the wider human economy [1]. Global data reveal 1.81 billion people live in locations that have been exposed to significant risks of inundation to depths greater than 0.15 m over the last 1–100 years [2]. The top ten subnational regions worldwide with the greatest relative exposure to flood risk are in Africa and Asia. Jonglei and Unity States in South Sudan are among the regions at the greatest risk of flooding. Previous studies demonstrate that these flood risks are projected to increase under future climate warming [3].

South Sudan faces several challenges in terms of water management. A lack of infrastructure, inadequate management practices, and ongoing conflict have all contributed to frequent flooding and droughts [4]. For example, in July 2019, severe flooding left more than 908,000 individuals homeless and prompted the government to declare a state of emergency [5]. Additionally, in October 2020, the United Nations Office for the Coordination of Humanitarian Affairs estimated that 800,000 people were affected by flooding along the White Nile, with almost 366,000 displaced. In 2022, the UN Refugee Agency reported that 900,000 people were affected by flooding over two-thirds of the country.

Despite the extensive damage and losses caused by floods, insufficient data are available to reduce the flood risk and aid adaptation efforts. To date, no scholars have engaged in a comprehensive synthesis of existing information on flood risk patterns and the root causes thereof in South Sudan. Trend analysis is commonly used to effectively understand historical flood changes and can play key roles in the fields of finance, economic healthcare, and environmental monitoring [6]. This kind of analysis is used to identify patterns and tendencies in data derived over time.

Many scholars worldwide have evaluated meteorological, hydrological, and climatological variables using the Mann–Kendall (MK), modified Mann–Kendall (mMK), and Spearman rho tests as well as simple linear regression analysis (LRA) [7]. However, these tests require restrictive assumptions: the time series are independently structured, data distribution is normal, and the data length is considerable. They cannot identify how low and high values affect the trends [8]. Therefore, ref. [9] developed an innovative trends analysis (ITA) technique to overcome these challenges, which is now widely used to detect trends among hydrological variables. For example, ref. [10] used ITA to analyze annual maximum rainfall series, and ref. [11] used it to investigate the trends of monthly mean stream flows in the Black Sea of Turkey. Ref. [12] used MK, mMK, and ITA approaches to investigate long-term trends and spatial variability in rainfall over southern Bangladesh. Ref. [13] used ITA to analyze rainfall anomalies in southern Italy. Many studies have compared the MK, Sen’s slope estimator (SS), LRA, and ITA and have confirmed that ITA is superior to traditional trend-detecting tests. Now, trends are commonly explored using all of the MK, mMK, and ITA approaches [14,15].

However, only a few researchers have used newly developed ITA techniques to explore how rainfall differences and river flows affect frequent extreme flooding. Even fewer have employed this technique to investigate the non-monotonic and/or hidden trends in the Expert Teams on Climate Change Detection and Indices (ETCCDI) rainfall dataset or stream flows when attempting to define the principal factors contributing to extreme flooding on the basin scale.

Most research conducted in South Sudan has focused on broader societal issues [16,17,18]. It remains unclear how local rainfall patterns and streamflow dynamics contribute to the increasing frequency of floods in the region. Therefore, this study explored how hydroclimatic conditions in Jonglei and Unity States were affected by annual total rainfall, the frequency and intensity of extreme rainfall events, and the stream flow of the Nile River. The ITA approach was utilized to investigate precipitation patterns and clarify how these affected flood frequency in both basins.

Study Areas

South Sudan (Figure 1a) is the 54th independent country of Africa and the 193rd member of the United Nations. The East African country is landlocked, and its physical features are defined by the Nile River that runs from south to north across the center of the country [19]. The 2008 census of the National Bureau of Statistics revealed that South Sudan was home to a population of 12 million. The climate is tropical and hot all year around, but somewhat less so in the rainy summer, when the humidity is high. More than 50% of all annual rain falls from June to August, peaking in July and August [20]. The country is divided into 10 states, of which Jonglei and Unity States are most severely affected by floods.

Jonglei State (Figure 1b) is the largest state in South Sudan (approximately 122,581 km²), and its Bor County is most affected by floods given its proximity to the Nile River [21]. According to the National Bureau of Statistics, Bor County has a total area of 13,931 km² and a population of 305,305. The predominant topography is flatland. The climate is tropical wet/dry savanna, with a maximum annual temperature of 43.0 °C and a minimum of 31.0 °C, and the annual precipitation is 800–1600 mm.

The Bentiu and Rubkona regions of Unity State (Figure 1c) are also affected by flooding. According to the National Bureau of Statistics, Unity State has an area of 37,836 km² and a population of 824,700. It has a subtropical steppe climate with an average temperature of 30.5 °C, and the annual average precipitation is 500 mm.

2. Materials and Methods

2.1. Datasets

South Sudan is characterized by a high degree of socio-economic fragility and weak institutional and human capacities. A civil war has raged for more than five decades: political and security systems are fragile, and economic hardship is prevalent. Investment in infrastructure is slow, and hydrological and meteorological stations destroyed during the civil war have not been rehabilitated. Meteorological coverage is also poor, so this study utilized hydrometeorological datasets derived via satellite-based remote sensing.

2.1.1. Rainfall

Analyses drew from the quasi-global (50° S to 50° N) Climate Hazards Group InfraRed Precipitation with Station (CHIRPS) rainfall dataset. These data are available from 1981 to the near-present. CHIRPS features high-quality daily rainfall climatological data derived via satellite imagery at a resolution of 0.25°. In-situ station data are used to create gridded rainfall time series that aid trend analysis and seasonal drought and flood monitoring. CHIRPS data are particularly robust for Ethiopia, Kenya, and Tanzania. The correlations are high, and the errors and biases low at both the decadal and monthly time scales [22]. CHIRPS data from East Africa have been evaluated at daily, decadal, and monthly time scales by comparing the data to rain gauge findings from about 1200 stations [23]. CHIRPS products have also been compared to those of two similar, operational satellite rainfall systems: the African Rainfall Climatology version 2 (ARC2) and the Tropical Applications of Meteorology using SATellite data (TAMSAT). CHIRPS was significantly better than ARC2, with higher skill and low or no bias. It also performs somewhat better than the latest versions of TAMSAT data at both decadal and monthly time scales.

In the present study, CHIRPS data were correlated with monthly rainfall levels recorded at the Bor and Bentiu meteorological stations in Jonglei and Unity States, respectively. The correlation coefficients were satisfactory (0.82 and 0.81, respectively), the Nash-Sutcliffe efficiencies were 0.67 and 0.61, and the root mean square errors were 38.3 and 42.2 mm/month, respectively.

The Expert Team on Climate Change Detection and Indices (ETCCDI) was used to derive the annual total precipitations on wet days (PRCPTOTs), the simple precipitation intensity indices (SDIIs), and the annual number of days with precipitation ≥ 20 mm (R20). These indices were calculated using Formulas (1), (2), and (3), respectively:

$$\mathrm{PRCPTOT}={\sum }_{\mathrm{i}=1}^{\mathrm{I}}{\mathrm{R}\mathrm{R}}_{\mathrm{i}\mathrm{j}}$$

(1)

where RR_ij is the precipitation on day I in period j, and I is the number of days in j.

$$\mathrm{SDII}={\displaystyle \frac{\sum _{\mathrm{w}=1}^{\mathrm{w}}{\mathrm{R}\mathrm{R}}_{\mathrm{w}\mathrm{i}}}{\mathrm{w}}}$$

(2)

where RR_wi is the daily precipitation on wet days in period j, and W is the number of wet days in j.

RR_ij = ≥ 20 mm

(3)

where RR_ij is the daily precipitation on day (i) in period (j)

2.1.2. River Flow

Trend analyses focused on the observed river flows at Mangala station (Figure 2) as recorded by the Ministry of Water Resources and Irrigation. Mangala station was established in 1905 to record the water levels of the Nile River. However, the present river flow dataset includes only observations from 2008 to the present.

Moreover, more than 20% of all data are missing because the civil war and political instability compromised station operation. To compensate for the lack of long-term continuous river flow data from Mongala, water level data from Lake Victoria (the source of the Nile River) were used, as the Lake Victoria outflow significantly affects river flow fluctuations at Mongala. Lake outflow has been modeled previously using lake water levels [24]. Lake Victoria water level data were obtained from G-REALM (https://ipad.fas.usda.gov/cropexplorer/global_reservoir/, accessed on 30 April 2025), which is a global dataset of lake and river water levels derived via satellite altimetry.

2.1.3. Flood Area

South Sudan has grappled with the dual challenges of flooding and drought for many decades, leading to widespread displacement, poverty, and food insecurity. Previous studies emphasized the instances of floods and droughts since 1900, highlighting a significant increase in flood occurrences [25]. Despite the country’s rich water resources, effective water management continues to be a challenge, primarily due to insufficient storage facilities and irrigation systems.

Figure 1 presents flooding from 2019–2024. The datasets and maps were generated by the United Nations Satellite Center (UNOSAT), which processes VIIRS-NOAA imagery using various methods. All satellite imagery was thoroughly reviewed, often via comparison of multiple images to detect significant changes. Several automated remote sensing techniques were employed to extract flood and land cover data, and these were subsequently reviewed and refined. These flood maps help clarify how land cover and precipitation affect flooding. We used the Global Surface Water dataset to explore long-term changes in flooding in Jonglei and Unity States.

2.2. Innovative Trend Analysis

Traditional trend analysis methods such as the MK and mMK systems detect monotonic trends but cannot consider variations at the different data percentiles within a single computation process [26]. The new ITA method improves hydrometeorological analysis: this robust trend detection technique has been used to study rainfall [27] and temperature [28] variations worldwide.

ITA first divides monthly mean rainfalls into two subseries, (X = (x₁, …, x_n/2) and (Y = (x_n/2+1,…, x_n), with an equal number of observations. Next, both subseries are arranged in ascending order and plotted against each other using a Cartesian coordinate system. The first half of each subseries is plotted on the x-axis and the second half on the y-axis. A straight line is then fitted to the scattered plot. This reveals either a monotonic trend or no trend. If the scatter points are clustered above the 45° line (1:1), the trend is increasing. If the points are clustered below the 1:1 line, the trend is decreasing. If the points are concentrated along the trend line, there is no trend. When calculating the trend indicator, the following equation is used:

$$\mathrm{B}={\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\displaystyle \frac{10\left(\mathrm{x}\mathrm{j}-\mathrm{x}\mathrm{k}\right)}{\overline{\mathrm{x}}}}$$

(4)

where B is the ITA slope, n is the number of observations in each subseries, xj and xk are the values of the consecutive subseries, and $\overline{x}$ represents the mean of the subseries (xk). A positive B value indicates an increasing trend, and a negative B value indicates a decreasing trend. Unlike the most commonly used classical methods of trend analysis, the ITA is free from assumptions such as serial correlation, non-normality, and length of the record [29]. Moreover, by using the ITA method, significant subseries trends (sub-trends) can be observed from the graphical representations [30].

The ITA was used to explore the following hydrometeorological elements and revealed characteristic long-term trends in all elements.

• PRCPTOT, R20 mm, and SDII values of the spatial average precipitations over the entire Sudd wetland and Jonglei and Unity States;
• PRCPTOT, R20 mm, and SDII values of the precipitations in each CHIRPS grid (0.25° resolution) within Jonglei and Unity States;
• River flows at the Mangala hydrological station and Lake Victoria water levels;
• PRCPTOT, R20 mm, and SDII values for the spatial average precipitations in the Mangala basin and the Lake Victoria catchment.

It is noted that the data periods for CHIRPS precipitation, Mangala river flow, and Lake Victoria water levels used in the above ITA analysis differed. CHIRPS precipitation data spans from 1981 to 2023, whereas the other datasets cover 2008 to 2024. An overlap between CHIRPS precipitation and Lake Victoria water levels data exists from 1994 to 2023. The analysis results for Mangala river flow reflect recent trends only. By using Lake Victoria water level as a proxy, we aimed to capture longer trends in the river flow from the upstream area. Further, by analyzing long-term precipitation trends in the upper catchment (Victoria Lake basin and Mangala’s upper catchment), we examine the relationship between flood conditions in the target region (Sudd region: Jonglei and Unity State) and both local precipitation and upstream influences.

Additionally, missing flow data at Mangala were estimated using water level–flow rate (H-Q) curves derived from observed river water levels.

3. Results

3.1. Inter-Annual Fluctuations and ETCCDI Interrelationships

Scholars have employed various indices of climatic variability and extremes for many years in efforts to clarify the evolving patterns of extreme weather on a global scale. Some of these indices have been internationally coordinated, with considerable advances made in recent years [31].

The CHIRPS daily gridded rainfall datasets for Jonglei and Unity States, which include more than 40 years of records, were analyzed to investigate trends in annual rainfall. Analysis included three precipitation-based indices of the ETCCDI: the PRCPTOT, R20 mm, and SDII, which are assessed annually (Figure 3 and Figure 4). Note that in Jonglei State, the PRCPTOT decreased in recent years (Figure 3a). The R20 mm varied from year to year but declined in recent years (Figure 3b). The SDII varied over time but declined in recent years (Figure 3c). All of the PRCPTOT, R20 mm, and SDII fell sharply from 2009 to 2010.

In Unity State, PRCPTOT also varied from year to year, decreasing sharply before 2010 and declining further in recent years (Figure 4a). The R20 mm also varied in previous years, but fell markedly from 1985 to 2010, and then fell further in later years (Figure 4b), The SDII was inconsistent over time but declined in recent years (Figure 4c).

Figure 5 presents histograms and scatter plots of PRCPTOT, R20 mm, and SDII for both Jonglei and Unity States. The graph offers a clear visual narrative, explaining the complex relationship among these indices. Examination of the plots affords valuable insights into index patterns and interactions. For example, the precipitation histogram pattern in Jonglei State is especially remarkable. In recent years, the annual total precipitation during wet days has been about 900 mm. However, the R20 mm values are skewed to the right, being predominantly between 1 to 7 days, with a few outliers. Further investigation of the precipitation intensity reveals that the SDII exhibits an exceptional distribution, being both remarkably balanced and symmetric, with most data clustered around the center. This suggests that precipitation could be accurately predicted and was evenly distributed across the observed period. The R20 mm and SDII values were strongly correlated (coefficient 0.86). Correlations were weaker between the R20 mm and PRCPTOT values and between the SDII and PRCPTOT values, at 0.59 and 0.47, respectively.

For Unity State, the PRCPTOT distribution was left-skewed with no outliers, and the range was narrow (600–1000 mm). This finding suggests that in recent years, the annual total precipitation on wet days was consistently about 600 mm. In contrast, the R20 mm value exhibited a right skew, with values between 1 and 8 days, with some noticeable outliers. The SDII evidences a unimodal symmetric distribution, with all data concentrated around the center, reflecting a consistent precipitation intensity from 1981 to 2023. The strong correlation (0.82) between the R20 mm and SDII values suggests that as the frequency of rainy days changed, so did the rainfall intensity. The moderate strength correlation (0.51) between the R20 mm and PRCPTOT values, along with the moderate strength correlation (0.73) between the SDII and PRCPTOT values, offer a useful context when evaluating the broader implications of precipitation trends.

Figure 6 illustrates the relationship between R20 mm and SDII over time. Earlier observations revealed that the two indices shared strong correlations of 0.86 and 0.82 for Jonglei and Unity States, respectively. Mapping this relationship through time afforded valuable insights into the dynamic patterns of rainfall that enhanced understanding of local climatic behaviors. As shown in Figure 7a,b, the R20 mm peaked at a very high value in the 1980s before decreasing in subsequent years. This downward path was mirrored by both the SDII and R20 mm values, highlighting a concerning shift in the climatic patterns of the two regions.

3.2. ITA Results in Terms of the ETCCDI and Spatial Distribution

3.2.1. PRCPTOT, R20 mm, and SDII Data Reveal the Spatial Average Precipitations over the Entire Sudd Wetland and Jonglei and Unity States

Figure 7, Figure 8 and Figure 9 present the results of the ITA, which involved examination of more than 100 grid points of the Sudd wetland and Jonglei and Unity States.

In the Sudd wetland (Figure 7a), an initial (but non-significant) decreasing PRCPTOT trend was noted in the first half of the studied time period, followed by no trend.

A significant, consistent monotonic decreasing trend of R20 mm (Figure 7b) was apparent in the first half of the period. However, in the second half, the trend in SDII (Figure 7c) exhibited a monotonic decreasing trend that was initially significant, followed by a non-significant decrease. In Jonglei State, a monotonic, non-significant increasing trend in PRCPTOT was apparent in the first half of the period (Figure 8a).

In contrast, the R20 mm evidenced a more intense and substantial decreasing trend in the first half of the period (Figure 8b). However, the decreasing SDII trend in the first half of the period was not significant (Figure 8c). Similarly, in Unity State, a consistent, monotonic decreasing trend in PRCPTOT was noted in the first half of the period (Figure 9a), as was a significantly decreasing trend in R20 mm (Figure 9b). However, the decreasing SDII trend in the first half of the period was not significant (Figure 9c).

The spatial distributions (Figure 10a) revealed decreasing PRCPTOT trends in southern and central Jonglei State with strong gradients ranging from –2.3 to –0.26 mm/year but increasing trends with gradients ranging from 1.9 to 7.9 in some northern parts of the state.

Decreases were also observed in R20 mm (Figure 10b) and SDII (Figure 10c) trends, with gradients ranging from −0.17 to −0.03 mm/year and −0.06 to −0.018, respectively, in the central, southern, and northern parts of the state. In Unity State, the data revealed a decreasing PRCPTOT trend (Figure 11a), with a strong gradient from −2.6 to −0.37 mm/year throughout the state, albeit with a weaker slope of 0.19 in the southern part. Both R20 mm (Figure 11b) and SDII (Figure 11c) also evidenced declining trends, with gradients ranging from −0.13 to −0.057 mm/year and −0.039 to −0.012, respectively, across the state.

As shown in Figure 8 and Figure 9, ITA trends for areal averaged precipitation in Jonglei and Unity were mostly not statistically significant, except R20 mm in Unity. However, we observed a significant trend at certain grid points within each region (e.g., significant decreasing trend in the first half at the grid point with regional maximum ITA slope for R20 mm and SDII in Jonglei, for PRCTOT in Unity).

3.2.2. River Flow at the Mangala Hydrological Station and Lake Victoria Water Levels

River flow at the Mangala station, as measured by the Ministry of Water Resources and Irrigation from 2008 to 2024, was subjected to trend analysis. Figure 12a presents the annual monthly river flow at Mangala station from 2008 to 2024. The highest flows were observed between September and November and the lowest from March to June. The annual mean flow was 317,470 m³/s, with a standard deviation of 176,181 m³/s. Mangala station is situated in central South Sudan. The rainy season in northern South Sudan typically lasts from May to October, while the central and western regions generally experience their first rains from April to June, followed by a second rainy season from August to November [32]. The annual ITA of the river flow at Mangala station revealed a clearly increasing trend (confidence level 95%). The magnitude of the ITA change was 1712 and the standard deviation of the slope was 295 (Figure 12b).

Figure 12c presents the ITA data for Lake Victoria water levels from 1994 to 2024. The level increased monotonically with a slope of 0.0314 and a standard deviation of 0.0025. The river discharge at Mangala station is affected by rising water levels in Lake Victoria. Figure 13 presents the trend analysis for the Mangala watershed. No PRCPTOT trend was initially observed (Figure 13a), followed by a significantly increasing trend in the second half of the period, consistent with R20 mm values (Figure 13b). However, the trend in SDII (Figure 13c) decreased, albeit not significantly, in the first half of the study period, followed by no trend/a non-significant increase in the second half.

3.2.3. PRCPTOT, R20 mm, and SDII Values for the Spatial Average Precipitations in the Mangala Basin and the Lake Victoria Catchment

Similarly, in the Lake Victoria watershed (Figure 14), PRCPTOT exhibited a significant increasing trend (Figure 14a) in the second half of the study period, combined with a significant, consistent, monotonic increasing trend in R20 mm (Figure 14b). However, the increase in the SDII (Figure 14c) was not significant.

4. Discussion

Analysis of rainfall patterns revealed a significant decrease in total annual precipitation, and also notable decreases in both the frequency and intensity of rainfall, in Jonglei and Unity States. However, such trends were in absolute contrast to observations made at Mangala station, where river discharges increased. Furthermore, the data contrast with the significantly increasing trends in annual total precipitation and the number of days with heavy precipitation in Lake Victoria and the upstream Mangala watersheds. The spatial distributional trends of PRCPTOT, R20 mm, and SDII decreased in most areas of central and southern Jonglei State, but the northern regions exhibited both increasing and decreasing trends. In contrast, in Unity State, the spatial distributions of PRCPTOT, R20 mm, and SDII consistently decreased over time, although to varying extents.

Flooding is typically triggered by extreme rainfall. However, in the context of South Sudan, the situation is exacerbated by intense rainfall in the Lake Victoria basin. This causes the water levels of Lake Victoria to rise significantly, increasing the water flow into South Sudan. The subsequent increase in river discharges poses challenges; South Sudan cannot control such influxes, so the area susceptible to flooding has significantly increased (Figure 15), as revealed by the global surface water dataset. The area affected by surface water rose markedly in 2014, followed by a decline in 2015 and then a renewed increase mid-2018. Ever since that time, surface water areas have increased in both Unity and Jonglei States, highlighting the substantial impact of rising Lake Victoria water levels on river flow discharges in South Sudan. Together, these dynamics suggest that localized rainfall patterns are not the primary drivers of flooding in Jonglei and Unity States. Rather, flooding is essentially influenced by external factors, specifically upstream rainfall in the Lake Victoria basin. Flooding overwhelms the local infrastructure in South Sudan, especially Jonglei and Unity States. Effective mitigation of flooding requires understanding the effects of rainfall and upstream basin inflows on downstream flood events, and the use of both physical and conventional hydrological models enables a comprehensive assessment of the flow dynamics and the characteristics of water from upstream sources. These models simulate various scenarios when quantifying the contributions of upstream rainfall and discharges to downstream flooding.

This finding highlights the significant socioeconomic implications of flooding in South Sudan and informs flood prevention policies. It underscores the need to redirect policy efforts towards regional cooperation, improve forecasting capabilities, and develop infrastructure that addresses external drivers of flooding. Furthermore, it emphasizes the importance of providing sustained support to communities adversely affected by such events. By integrating real-time data with predictive modeling, decision makers can establish effective early warning systems, enhance flood forecasting, and implement targeted mitigation strategies such as floodplain zoning and infrastructure planning to strengthen resilience and minimize the impacts of recurrent flooding.

5. Conclusions

This study employed advanced trend analysis to investigate the rainfall patterns in Jonglei and Unity States and their upstream basins. Analyses drew from a CHIRPS rainfall dataset with a resolution of 0.25° and revealed decreases in PRCPTOT, R20 mm, and SDII metrics in both Jonglei and Unity States but increases in the upstream basins. No substantial evidence suggests that annual rainfall in Jonglei and Unity States is significantly increasing over time.

Trend analyses of river flow at Mangala station, and the water levels in the Lake Victoria and Mangala watersheds, revealed clearly increasing trends (confidence levels 95%). Global surface water data indicated a rise commencing in mid-2018. This finding suggests that river flow is significantly influenced by increasing water levels in Lake Victoria. Notably, the rises in both the water surface area and river flow conflict with the observed decreases in rainfall in both Jonglei and Unity States. Local rainfall is not the primary cause of frequent flooding. This study is subject to data gaps and methodological limitations. The Sudd wetland experiences substantial data gaps due to disruptions in the gauge network for measuring rainfall, river stage, and evaporation caused by civil war. Many monitoring stations ceased operations following the civil wars from 1983 to 2005 and again after 2013, resulting in discontinuous discharge series along the Nile River. Furthermore, the absence of an internally derived significance test in the original ITA, which is primarily descriptive, means that analysts may misinterpret any deviation from the 1:1 line as “significant.” This can inflate false positives in data-sparse contexts.

Given these findings and constraints, further investigation and management of key upstream issues will be essential. In addition, we plan to use a hydrological model to examine not only long-term changes in precipitation in the upper reaches but also the effects of land use changes and human impacts in the region on inflow to the Sudd region.