Utilizing the NDWI and CCM with Heatmap Analysis Waterbody Fluctuations in Al-Hawizeh Transboundary Marshes: A Strategy for Supporting Decision-Makers in Achieving Sustainability

Wetland ecosystems are among the most biologically productive and diverse habitats on our planet, serving as hubs for numerous ecosystem services that benefit both the environment and human societies. The most important of these services are water filtration, flood regulation, bank stabilization, carbon storage, the provision of habitats for a vast array of plant and animal species [1], [2]. Wetlands are estimated to cover only 6% of the total land area, yet they are the primary ecological home of over 40% of the global species underscoring their ecological significance [3]. Despite their crucial role, Wetlands are among the most fragile ecosystems on Earth. Since the early 20th century, it is estimated that approximately 70% of wetlands worldwide have disappeared [4], [5]. The persistence and protection of these critical ecosystems depend on the success of water management, which requires the timely collection of accurate and reliable information regarding the spatial extent and temporal dynamics of water distribution for wetland planning.

Because of the extensive, unique, and inaccessible nature of wetlands, the conventional field monitoring methods are often impractical for extensive areas. In this context, remote sensing technologies have emerged as indispensable tools [6], [7], [8].

Satellite remote sensing is considered a cost-effective, environmentally friendly, and advanced method for monitoring soil water saturation over extended periods [9]. Given the importance of spectral indices derived from multispectral imagery in accurately mapping and monitoring water bodies and their variations, these indices provide essential information to water resource managers and decision-makers [10], [11], [12], [13].

Accordingly, the study employed these indices to analyze the marshland regions of southern Iraq, with particular focus on the Al-Hawizeh Marshes.

In the 1990s, a large-scale engineering project was carried out to drain the marshes, which reduced their area by more than 90%. This has been described as one of the most severe human-made environmental disasters of the twentieth century [14], [15], [16].

The marshes of southern Iraq, including Al-Hawizeh Marshes, experienced significant natural decline and a profound ecological stress in the late twentieth century. However, the collapse of these ecosystems was not solely a natural phenomenon. In the 1990s, a large-scale engineering project was implemented to drain the marshes, reducing their area by more than 90%. This event has been described as one of the most sever human-made environmental disasters of the twentieth century [14], [17], [18].

Similarly, the importance of the Al-Hawizeh Marshes lies in their designation as a Ramsar site and their inclusion on the United Nations Educational, Scientific and Cultural Organization (UNESCO) World Heritage List. This significance stems not only from their role as a habitat for numerous ecological species inhabiting shallow waters, but also from their remarkable resilience to recover after 2003, marked by the return of water to the marshland areas and the restoration of ecological diversity [6], [17], [19].

So, the aim of this article is to monitor the fragile environment of the Al-Hawizeh Marshes in order to determine whether they are undergoing an ecological period of recovery, which could restore their capacity to provide ecosystem services and sustain biodiversity, or whether they are experiencing hydrological instability leading to recession and desiccation, placing them in a state of ecological decline. Accordingly, this paper seeks to deliver a rigorous scientific assessment that contributes to understanding the future of this wetland and underscores the urgent need for water management strategies to support decision-making and achieve both environmental and cultural sustainability.

The analysis was conducted within the GEE environment Figure 1.

Step 1. Study Area: The official vector border was imported into GEE and defined as the Area of Interest (AOI).

Step 2. Satellite Data Preparation: Landsat 5, 7, and 8 Surface Reflectance collections were filtered for the years (2015, 2018, 2021 and 2024) and constrained to AOI.

Step 3. Composite Images: To mitigate data gaps (e.g., Landsat 7 SLC-off) and ensure consistency a daily composite (mosaic) image was generated.

Step 4. Water Detection: NDWI was calculated for each daily composite image NDWI result were used to create a binary mask (0 is land, 1 is water) to classify each pixel. The threshold of 0.2 is applied, which helps to distinguish water bodies and shallow waters, reducing errors caused by reflectance from vegetation or moist soil. The (QA_RADSAT) mask was used to exclude saturated data (e.g., highly bright areas such as sand or strong water reflections [19], [20].

Step 5. Calculate Water Area: Using the binary water masks, the total water area within the AOI was calculated as a percentage.

Step 6. Data Export: Time-series data of water area statistics were exported as a CSV files; derived raster maps were exported as GeoTIFF files.

Step 7. Analyze: Summary statistics (mean, minimum, maximum) were calculated for the water area and coverage metrics. Water permanence was analyzed seasonally and annually to provide deeper insights into the hydrological stability of the Marshes.

The focus is on Al-Hawizeh Marshes, a vast, transboundary wetland ecosystem situated along the border between southeastern Iraq and southwestern Iran [21]. Within Iraq, they extend approximately from (47$^{\circ}$25’ to 47$^{\circ}$45’) E in longitude, and from (31$^{\circ}$05’to 31$^{\circ}$40’) N in latitude [22]. It varies in size between summer and winter and from one year to another, this environment supports the local communities whose livelihoods (the Marsh inhabitants) depend on its natural resources [10]. Its hydrological system exhibits considerable complexity, as it is fed mainly by branches of the Tigris River in Iraq and the Karkheh River in Iran. The official boundaries of the protected area, whose total area amounts to 1,377 km$^2$ Figure 2.

This delineation represents the legally designated extent not correspond the actually covering water area because it was change between season and from year to year. That boundary of the protected AOI. The official boundaries of the protected area, whose total area amounts to 1,377 km$^2$.

The primary data source for this research was the collection of Landsat satellite imagery, specifically Tier one Surface reflectance products from Landsat5, Landsat7, and Landsat 8 in Thematic Mapper, Thematic Mapper Plus and Operational Land Imager, respectively. These collections provide atmospherically corrected data [23] the analysis was conducted for the entirety of the years 2015, 2018, 2021, and 2024 thus enhancing the accuracy of the following water detection analysis [24], [25].

The core of the water detection methodology was NDWI, proposed by Xu [26]. The spectral index is based on the difference in reflectance between water in the green and near-infrared (NIR) regions of the electromagnetic spectrum. Water bodies absorb NIR radiation significantly, while exhibiting strong reflection in the green band [26]. The index is calculated using the following formula:

For Landsat 8, the NDWI is calculated using (Band 3) and (Band 5), and for otherwise, it is (Band 2−Band 4) [27].

Although there is a variety of advanced indices such as the Modified NDWI (MNDWI) [28]; this index was adopted for computational efficiency, and demonstrated performance in delineating open water bodies across vast vegetated wetland landscapes such as Al-Hawizeh Marshes [16].

The analysis yielded detailed quantitative and visual results, revealing significant and dramatic fluctuations in the waterbody area of Al-Hawizeh Marshes across the selected years. The findings are presented in a comparative data table that summarizes the mean monthly water area.

This result reveals a significant variation in the Marshes hydrological condition from year to year, as seen in Table 1.

In contrast, years 2018 and 2024 exhibit moderate, significantly desiccated conditions, with annual average water areas of only 48.4 km$^2$ and 49.6 km$^2$, respectively. The year 2015 represents a more moderate, baseline condition, with an average water area of 83.2 km$^2$. With an average annual water area of 448.4 km$^2$, year 2021 stands out as an extraordinarily wet year. This extreme fluctuation between extensive inundation and severe dryness underscores the dynamic and vulnerable nature of the marsh’s ecosystem.

To substantiate this, each year will be analyzed separately using GEE maps and statistical analysis as listed in table separately.

In 2015, Al-Hawizeh Marshes exhibited moderate water levels with distinct seasonal variation, representing a baseline condition for this study. The annual average water area was approximately 83.2 km$^2$. A time-series chart for 2015 Figure 3 and the water distribution map as shown in Figure 4, confirm a typical hydrological pattern for the region.

The graph Figure 3 illustrates how the average water area changed during 2015. It exhibits a clear peak in spring (April) at 163.5 km$^2$, followed by decrease to its lowest point in late Summer/early Autumn (September) at 40.3 km$^2$, before rebounding a recovery towards the end of the year.

The year 2018 marked a period of significant dryness, with the annual average water area plummeted to 48.4 km$^2$. This is illustrated in the monthly average chart (Figure 5) and the spatial distribution map for 2018 (Figure 6).

This chart Figure 5 shows a peak water area in March (95.9 km$^2$), followed by a decline that accelerates dramatically in the second half of the year. By the end of the year, the water area falls to a lower value.

The map Figure 6 for 2018 shows a substantial reduction in water coverage compared to 2015. Large expanses are classified as “no water”, dominated by yellow, with only small, fragmented pockets of “high water” visible, highlighting a critical dry period.

In stark contrast to the rest of study period, 2021 was an exceptionally wet year. The average water area surged 448.4 km²,around ten times the 2018 average. This is clearly illustrated in the monthly average chart Figure 7, and the water distribution map Figure 8.

Figure 7 represents a dramatic reversal from the conditions observed in 2018. The monthly chart Figure 7 shows consistently high-water values throughout the year, peaking in March at 629.1 km$^2$ and remaining above 200 km$^2$ even at its lowest point in October. This suggests a major flood event or a period of sustained, unusually high-water supply to the marsh system. Furthermore, the chart for 2021 depicts water area on a completely different scale from the other years. Levels remain exceptionally high for the entire year, peaking in March and gradually decreasing, but never approaching the low levels recorded in other years. It should be noted that the chart in Figure 7 differs from charts in Figure 3, Figure 5, and Figure 9, where the y axis values do not exceed 170 km$^2$ and are divided into intervals of 10 km$^2$. But, the y axis in the year 2021 rises to 650 km$^2$, with the lowest values reaching 200 km$^2$, and the divisions set at 50 km$^2$. This distinction reflects the markedly higher water volumes observed in 2021, which is characterized as a flood year with generally elevated water levels.

This map Figure 8 illustrates the extensive flooding in 2021. Most of the AOI is covered by water, areas of “high water” (blue), representing a dramatic reversal from the conditions seen in 2018.

The interpretation of 2021 reveals the emergence of large volumes of water, as shown by statistical and spatial analysis. This is attributed to the fact that it followed the year of the COVID-19 pandemic, which allowed the environment to recover. One of the most prominent reasons was the increase in rainfall and the inflow of water from upstream sources. It also reflects the policy of ineffective water management across Iraq.

In 2024, the marsh conditions appeared to revert to a dry state, with hydrological patterns closely resembling those of 2018. The annual average water area was 49.6 km$^2$, almost identical to the 2018 average. The time-series chart Figure 9 and map Figure 10 show that fluctuation.

Time-series chart Figure 9 reveals a notable spring peak in April (138.8 km$^2$), which was surprisingly higher than the peaks in 2015 or 2018. However, this wet period was short-lived and was followed by a rapid decline to very low levels in late summer (15.2 km$^2$ in August).

The 2024 map Figure 10 shows a landscape largely devoid of surface water, closely resembling the 2018 map. It is again dominated by “no water” (yellow) areas, with only sparse and fragmented water presence in blue. The overall condition in 2024 reflects a return to significant water scarcity after the extreme wet period of 2021.

Findings from the Al-Hawizeh Marshes between 2015 and 2024 reveal a system under extreme stress, characterized by profound inter-annual volatility with significant implications for its ecological functions and the human communities it supports. These results are summarized in Figure 11. As well, the results were further visualized through a heatmap, which highlights the intensity and density of waterbody values using color gradients, effectively illustrating the fluctuations revealed by the study (Figure 12).

The heatmap in Figure 12 illustrates that the dark blue color, representing higher waterbody values, reached approximately 600 km$^2$ for about seven months during 2021. In contrast, it declined to lighter shades of blue, reaching the palest level during the driest months of 2018 and 2024. The correlation coefficient matrix (CCM), be displays in Table 2, which along with the accompanying heatmap, clarifies the linear correlations between the monthly waterbody area dynamics of Al-Hawizeh Marshes across the observed years (2015, 2018, 2021, 2024).

The interpretation in Table 2 of results based on Pairwise Correlation Analysis is as follows:

1. Years 2015 and 2018 ($r$ = +0.621): A moderate positive correlation indicates shared seasonal hydrological patterns, with both years exhibiting peak water areas during spring (March–April) and declines in autumn (October–December). However, the consistently lower water areas in 2018 (e.g., near-zero values in November–December) suggest intensified drought conditions relative to 2015, despite the structural similarity in seasonal cycles.

2. Year 2021 and Other Years ($r$ = -0.425 to -0.812): In 2021, there are moderate negative correlations with 2015 ($r$ = -0.425), and high negative correlations with 2018 ($r$ = -0.796) and 2024 ($r$ = -0.812).

3. This inverse relationship is attributable to the exceptional hydrological conditions in 2021, where mean water areas exceeded 500 km$^2$ during peak months (e.g., 629 km$^2$ in March). In contrast, 2018 and 2024 represent severe drought years (mean areas $<$ 100 km$^2$), while 2015 reflects intermediate conditions. The pronounced negative correlations underscore 2021 as an outlier driven by anomalous flooding, disrupting typical interannual coherence.

4. Years 2018 and 2024 ($r$ = +0.642): A robust positive correlation highlights the shared drought-driven hydrological regimes of these years. Both exhibit minimal water retention during winter months (e.g., 0 km$^2$ in December 2018; 5.96 km$^2$ in January 2024) and limited recovery in summer. This alignment suggests recurrent drought stressors, though 2024 shows marginally higher resilience (e.g., 138.8 km$^2$ in April 2024 vs. 73.8 km$^2$ in April 2018).

5.Years 2015 and 2024 ($r$ = +0.501): The moderate positive correlation reflects analogous seasonal phasing, with peak areas occurring in spring (April–June). However, the significantly reduced water extents in 2024 (e.g., 97.4 km$^2$ in May 2024 vs. 77.0 km$^2$ in May 2015) indicate progressive aridification, potentially linked to long-term climate stressors.

This strengthens the notion that, through spatial analysis and statistical assessment grounded in extracted indicators, it demonstrates that the Al-Hawizeh Marshes are in recession and are generally moving toward desiccation.

The methodology of this study, which relies on the GEE platform and Landsat satellite data, has proven highly effective in monitoring the Al-Hawizeh Marshes ecosystem during the period between 2015 and 2024. This large and complex system is characterized by extreme hydrological fluctuations, with the marshes experiencing severe drought in 2018 and 2024, contrasted by extensive flooding in 2021.

The results underscore the critical state of the marshes and the urgent need for cooperative transboundary water management. Sustainable and equitable water sharing that ensures environmental flow requirements is essential for the marshes’ long-term existence, their distinctive biodiversity, and preservation of ancient civilization cultural sustain.

Preserving their sustainability necessitates a comprehensive approach that takes into account crucial factors:

Firstly, preserving environmental flow: ensuring sufficient water inflows throughout the year to sustain the marshes distinctive ecology. Secondly, regional cooperation: recognizing the transboundary nature of these essential resources. Finally, community engagement: supporting residents’ livelihoods and empowering them to take part in decision-making and resource management.

To ensure the ability of this fragile ecological fractal system to endure in the long term, adaptive planning that adopts resilient and stainable strategies to deal with drought and floods together represents a rational methodology to water management, which contribute to supporting the decision-maker and to achievement of sustainability.