Environmental progress and economic growth: interplay and implications for poverty alleviation in Bangladesh

The Normalized Difference Vegetation Index (NDVI) is a widely used remote sensing metric that provides a reliable estimate of vegetation cover and health (Yengoh et al., 2016). NDVI has been used extensively to monitor land use changes, agricultural productivity, and the effects of climate change (Huete et al., 2002). Given the crucial role of vegetation in supporting biodiversity, regulating the climate, and sustaining livelihoods, understanding the impact of socio-economic policies on NDVI is essential for sustainable development strategies. With its exceptional access to satellite imagery and sophisticated computational capabilities, Google Earth Engine (GEE) has become a potent platform for large-scale environmental investigation (Gorelick et al., 2017). GEE plays a crucial role in our research, allowing us to extensively assess NDVI trends over long periods and broad geographic regions. Its exceptional access to satellite imagery and sophisticated computational capabilities make it a perfect tool for researching how socioeconomic policies in Bangladesh affect the environment (Gorelick et al., 2017; Kumar and Mutanga, 2018; Tamiminia et al., 2020; Yang et al., 2022; Velastegui-Montoya et al., 2023).

Bangladesh is among the countries most heavily impacted by climate change, despite having little contribution to global emissions of greenhouse gases (Ali, 1996; Rahman et al., 2019). The effects of global climate change are not a distant concern for Bangladesh but an urgent and pressing issue. The country, despite its defenseless position, has been negatively impacted by climate change, leading to significant loss of life and wealth each year. The frequency of natural disasters, including flooding, landslides, storms, cyclones, droughts, erosion, and salinization, has increased due to climate change, making it a matter of immediate concern for the nation (Ali, 1996; Huq et al., 1999; Huq, 2001; Biswas, 2013; Dastagir, 2015; Das et al., 2024). Because of its location, topography, and socioeconomic status, the country faces substantial challenges to global climate change. Environmental Justice Foundation reported that by 2050, one in seven Bangladeshis is expected to be displaced by climate change, and the country may lose around 11% of its land. This would impact the estimated 15 million people who live in the low-lying coastal region (Changkakati, 2022).

Using satellite imagery to monitor and identify forest areas, numerous researchers have concentrated on reducing the effects of climate change and establishing sustainable forestry (Verbesselt et al., 2012; Menaka et al., 2013; Wang et al., 2014; Arai et al., 2019). Recently, using the GEE application, several researchers have identified and visually assessed the changes in forest land and degradation of numerous countries with NDVI (Arai et al., 2019; Jena and Pradhan, 2019; Zhang et al., 2019). Using satellite images, a few scholars evaluated the state of the forests in real-time and tracked changes in deforestation (Rao et al., 2005; Bhandari et al., 2012; Verbesselt et al., 2012; Khan and Rahman, 2021). To determine the vegetation cover in Bangladesh, many researchers worked on mapping and classifying satellite images and presented a deforestation detection method that could operate in almost real-time (Yu et al., 2018; Changkakati, 2022; Das and Sarkar, 2023; Islam et al., 2024; Kanjin and Alam, 2024). Applying these advanced methods in Bangladesh has resulted in more successful forest conservation and management plans. By combining the NDVI with other vegetation indicators, researchers can better understand forest health’s geographical and temporal fluctuations. This can better inform policy decisions for sustainable forestry and climate change mitigation.

Poverty reduction policies in developing countries have become vital tools for promoting economic growth while addressing environmental sustainability. Many nations have demonstrated that well-designed strategies, such as social safety nets, microfinance initiatives, and green technology adoption, can uplift marginalized communities and improve environmental quality (Masud-All-Kamal and Saha, 2014; Habib and Jubb, 2015; Azman et al., 2022; Shah, 2024). For example, Ethiopia’s integration of clean energy solutions improved local livelihoods and reduced ecological degradation, while China’s ecological restoration projects in poverty-stricken regions enhanced vegetation cover and economic stability (Brown et al., 2011; Gardner et al., 2018; Guo and Liu, 2022). These examples highlight a growing recognition that poverty alleviation and environmental protection are interconnected goals, not competing priorities. However, the extent to which such synergies exist in other vulnerable regions, such as South Asia, remains understudied. Bangladesh, a climate-vulnerable nation, has implemented ambitious poverty alleviation programs since the early 2000s, including microfinance schemes, agricultural subsidies, and disaster-resilient infrastructure projects (Habib and Jubb, 2015; Shah, 2024; Hassan et al., 2025). While these efforts have lifted millions out of poverty, their environmental consequences particularly on vegetation cover and ecosystem health—are poorly understood. The country faces severe climate challenges, including intensified floods, cyclones, and soil salinity, threatening economic stability and natural resources (Ali, 1996; Huq et al., 1999; Biswas, 2013; Dastagir, 2015; Rahman et al., 2019; Das et al., 2024). Satellite-based monitoring tools like the NDVI have been used globally to assess vegetation health and land-use changes, offering insights into how human activities and policies impact ecosystems (Huete et al., 2002; Verbesselt et al., 2012; Menaka et al., 2013; Wang et al., 2014; Yengoh et al., 2016; Gorelick et al., 2017; Arai et al., 2019; Jena and Pradhan, 2019; Zhang et al., 2019). In Bangladesh, studies have mapped deforestation and agricultural shifts using NDVI, yet none have directly linked these trends to poverty reduction policies or economic indicators like GDP (Biswas, 2013; Yu et al., 2018; Das and Sarkar, 2023; Islam et al., 2024; Kanjin and Alam, 2024).

Globally, research on poverty-environment linkages reveals mixed outcomes. While initiatives in Madagascar and China show balanced economic growth and environmental recovery, poorly planned policies elsewhere have accelerated deforestation or soil degradation (Brown et al., 2011; Gardner et al., 2018; Wang et al., 2019; Fan et al., 2022; Guo and Liu, 2022). Economic growth (GDP) often correlates with environmental pressure, but evidence suggests that inclusive, nature-based strategies can mitigate this trade-off (Wang et al., 2019; Fan et al., 2022; Guo and Liu, 2022). In Bangladesh, satellite imagery analyzed through platforms like GEE has advanced real-time forest monitoring, but gaps persist in connecting these environmental datasets to socioeconomic policies (Gorelick et al., 2017; Kumar and Mutanga, 2018; Tamiminia et al., 2020; Yang et al., 2022; Velastegui-Montoya et al., 2023). Few studies explore how poverty programs influence vegetation health or whether economic gains align with ecological resilience. This study addresses these gaps by evaluating the impacts of Bangladesh’s poverty alleviation policies on NDVI trends and GDP across all administrative divisions from 2014 to 2023. Using GEE’s satellite imagery and computational tools, we analyze decade-long changes in vegetation health, linking them to regional economic growth and policy timelines. This research provides actionable insights for policymakers aiming to harmonize development and ecological conservation by uncovering trade-offs between poverty reduction and environmental sustainability. The findings will contribute to global discussions on achieving the Sustainable Development Goals (SDGs) in climate-vulnerable regions.

Bangladesh is located in South Asia, on the eastern side of the Indian subcontinent (20°34'–26°38' N, 88°01'–92°41' E). Bangladesh shares its western, northern, and eastern borders with India, while Myanmar lies to its southeast and the Bay of Bengal to its south. The total land area is approximately 147,570 km², with a coastline of around 580 km along the Bay of Bengal. Bangladesh has a tropical monsoon climate characterized by significant seasonal variations in rainfall. The year is divided into three distinct seasons: the warm season, the monsoon season, and the splendid season. The warm season spans from March to June, with average temperatures ranging from 25–35 °C. Influenced by the southwest monsoon, the rainy season spans from June to October, bringing heavy rainfall and humidity. The incredible season extends from November to February, with average temperatures ranging from 12–25 °C. Bangladesh is administratively divided into eight divisions: Dhaka, Chattogram, Khulna, Rajshahi, Barishal, Sylhet, Rangpur, and Mymensingh (Figure 1). Each division is subdivided into districts, upazilas (sub-districts), and unions. The country has a predominantly rural population, with about 65% living in rural areas. Agriculture is the mainstay of the rural economy, with rice, jute, tea, and sugarcane being the principal crops grown.

This study employed a rigorous and comprehensive methodology integrating remote sensing, geospatial analysis, and economic variables to analyze the relationship between NDVI values and economic growth rates (Figure 2). Using Landsat 8 imagery within GEE to calculate the NDVI from 2014 to 2023 (Table 1) and the subsequent spatial mapping and visualization of the NDVI data using ArcGIS Pro ensured the accuracy and reliability of the study’s findings.

Where NIR and red represent the near-infrared and red bands, respectively, which resulted in NDVI values that provided a comprehensive measure of vegetation health across Bangladesh over the specified period. Subsequently, the study used ArcGIS Pro for spatial mapping and visualization of the NDVI data, creating detailed maps that illustrated the spatial distribution and temporal changes in NDVI across different regions of Bangladesh. This facilitated a visual understanding of vegetation trends and helped identify areas with significant vegetation health dynamics over time. To complement the NDVI analysis, the study collected economic growth and population data from national surveys and the Bangladesh Bureau of Statistics (BBS), providing insights into regional economic trends, growth rates, and per capita income across various divisions of Bangladesh (Table 1).

To quantify temporal trends in GDP per capita and NDVI, linear regression was applied to estimate each region’s slopes over three periods (2014–2017, 2018–2020, and 2021–2023). The slope calculation follows the formula:

Where y represents GDP per capita or NDVI values, and x represents the corresponding years. A positive slope indicates an increasing trend, while a negative slope signifies a decline.

The study utilized Python to integrate and analyze the NDVI and economic data for statistical analysis. This study employed pandas, numpy, and matplotlib libraries to conduct correlation analyses and linear regression modeling. This allowed the author to understand trends, predict future values, and quantify the strength and direction of relationships between the variables (Stockemer, 2019). Additionally, variation patterns of GDP per capita and NDVI were categorized under four primary development models (Table 2): Integrated Sustainability (A), where both slopes are above zero; Growth-Oriented (B), where GDP per capita slope remains more significant than zero and NDVI slope remains below zero; Ecologically Driven Development (C), where NDVI slope remains above zero while GDP per capita slope remains below zero; and Deteriorative Development (D), where both slopes are below zero, indicating declines in both economic and environmental health (Fan et al., 2022).

This classification was facilitated by calculating the slopes using linear regression. The Pearson correlation coefficient was calculated using the formula further to assess the link between GDP growth and NDVI fluctuations.

Where n is cumulative years; i denotes the duration between 1 to n; $NDVIi$ is the NDVI value in year i; $GDPi$ is the GDP per capita in the year i; $NDVI̅$ is the normal yearly NDVI; and $GDP̅$ is the normal GDP per capita. The value of r_{NDVI, GDP} ranges from −1 to 1; if r_{NDVI, GDP} = 0, then there is no relationship between changes in the NDVI and GDP per capita; if r_{NDVI, GDP}<0, it shows that NDVI and GDP per capita exhibit divergent trends; if r_{NDVI, GDP} >0, it iindicates that NDVI and GDP per capita exhibit similar directional changes. The closer r_{NDVI, GDP} are to 1, means there is a strong relation between GDP per capita and NDVI variation.

The NDVI shows a declining trend for several divisions from 2014 to 2023 (Figure 3). NDVI assesses vegetation vitality and density, with higher values reflecting healthier, more robust vegetation.

The result indicates a consistent decline in NDVI values across all divisions, suggesting a widespread degradation in vegetation health over the past decade (Figure 4). While all divisions show a downward trend, Rangpur starts with the highest NDVI value and maintains relatively higher levels compared to other divisions, despite its decline. Khulna and Barisal also show a gradual decrease, whereas Chittagong exhibits one of the lowest NDVI values throughout the period.

Table 3 provides a comparative analysis of economic growth rates and corresponding NDVI changes across various Bangladesh divisions. It highlights the relationship between GDP growth rates and NDVI changes, offering insights into the dynamics between economic development and ecological health.

Table 3 presents the economic growth rates and corresponding NDVI changes for Bangladesh divisions between 2014 and 2023. The results show that all divisions experienced moderate and relatively similar average annual GDP growth rates (AAGR), ranging from 5.57% in Rangpur to 5.80% in Dhaka. Dhaka recorded the highest AAGR at 5.80%, indicating steady economic expansion in the country’s capital region. Chittagong and Khulna also demonstrated strong economic performances with AAGR values of 5.77% and 5.72%, respectively. Conversely, Rangpur had the lowest AAGR among the divisions, although the difference across divisions remains relatively small. However, despite this economic growth, all divisions reported a decline in NDVI values, suggesting a loss of vegetation health during the same period. The most significant decrease in NDVI was observed in Khulna (−0.061), followed closely by Barisal (−0.058) and Chittagong (−0.052). These regions, which showed strong economic growth, also experienced some of the most significant vegetation declines. In contrast, divisions such as Rajshahi (−0.039) and Mymensingh (−0.040) exhibited smaller NDVI reductions, although the negative trend was consistent across all areas.

Figure 5 illustrates the trends in economic growth rates across various divisions from 2014 to 2023, highlighting fluctuations and significant events that have shaped the economic landscape. Between 2014 and 2019, the economic growth rates for most divisions displayed a positive trend, typically ranging from approximately 6%–8%. This period reflects a phase of economic expansion, with the highest growth rates approaching 8% in divisions such as Rangpur and Dhaka. However, in 2020, all divisions experienced a sharp decline, with growth rates falling from around 4%–5%, likely due to the global economic impact of the COVID-19 crisis. After the downturn, a recovery started in 2021, with growth rates rebounding to around 6%–7%, although they did not quite reach the pre-2020 peaks. In the following years, 2022 and 2023, the growth rates stabilized or saw slight declines. Some areas, such as Khulna, experienced a more significant decrease, with growth dropping below 6%. This overall trend underscores these regions’ economic challenges and resilience as they navigate through periods of growth, crisis, and recovery.

The scatter plot and the regression line illustrate the relationship between the change in NDVI and GDP growth (Figure 6). The X-axis represents GDP growth in percentages, ranging from 3% to 8%, while the Y-axis shows NDVI changes, fluctuating from around −0.010 to slightly above 0.

The Pearson correlation coefficient of −0.1373, with a p-value of 0.2246, indicates a weak negative correlation between GDP growth and changes in NDVI. However, this relationship is not statistically significant because the p-value exceeds the 0.05 threshold. The regression line, expressed by the equation NDVI Change = −0.0005 * GDP Growth −0.0054, suggests that as GDP growth increases, NDVI change tends to become more negative, though only marginally. Data points are predominantly clustered near the lower end of the NDVI axis, indicating that most changes in NDVI are negative, irrespective of GDP growth. The slight downward trend of the regression line implies that higher GDP growth is modestly associated with a decline in vegetation health or density, although the effect is minimal. This trend could suggest that economic activities driving GDP growth might adversely affect vegetation. However, the weak correlation indicates the likelihood of other factors significantly influencing NDVI changes.

Similarly, the line graph depicting Bangladesh’s NDVI and Per Capita Income trends from 2014 to 2023 reveals distinct patterns (Figure 7). The X-axis spans from 2014 to 2023, and the Y-axis represents the Per Capita Income and NDVI values. The orange dashed line with “x” markers shows a consistent increase in per capita income over the years, with the regression equation Income = 79.11x + 157945.03 reflecting steady economic growth. On the other hand, the NDVI trend, represented by a yellow dotted line, remains relatively flat, hovering near zero on the Y-axis, indicating minimal variation or a slight decrease in NDVI over time. The NDVI equation, NDVI = −0.0078x + 16.5900, suggests a slight negative trend, hinting at a decline in vegetation health. While the rising per capita income indicates economic growth in Bangladesh, the near-stagnant NDVI trend suggests that this economic growth has not corresponded with improvements in vegetation health. Instead, it may point to potential environmental degradation or other factors negatively impacting vegetation despite the financial gains. In summary, the growth of Bangladesh’s economy, as evidenced by the increasing per capita income, has not necessarily led to positive changes in vegetation health, and there might even be a slight deterioration over time.

We investigated the relationship between GDP per capita and NDVI using a quadratic regression model, as illustrated in Figure 8. This regression framework is commonly employed to test the Environmental Kuznets Curve (EKC) hypothesis, which suggests that environmental quality tends to deteriorate in the early stages of economic growth. Still, after surpassing a certain income threshold, it begins to improve.

The derived regression equation is:

Here:

• b₂ = 2.641* 10^–7 represents the quadratic coefficient, indicating that NDVI decreases as GDP per capita increases initially.

• b₁ = −0.00116 represents the linear coefficient, suggesting that beyond a certain point, GDP per capita has a positive effect on NDVI.

• c = 0.735 is the intercept, the baseline NDVI when GDP per capita is zero.

From the equation, the turning point of the curve, where the effect of GDP per capita on NDVI shifts from negative to positive, can be calculated. The turning point occurs when:

This value represents the threshold level of GDP per capita at which the decline in NDVI slows, suggesting that environmental degradation may stabilize as economic development surpasses this point. The quadratic relationship observed in our analysis is consistent with the core idea behind the Environmental Kuznets Curve (EKC), which proposes that industrialization, deforestation, and urbanization can negatively impact the environment in the early stages of economic growth, leading to lower NDVI values. For our specific dataset covering the years from 2014 to 2023, we observe that in the early years, rising GDP per capita correlates with a decline in NDVI, likely due to economic pressures driving land-use changes that reduce vegetation cover. However, as GDP per capita approaches the threshold (around BDT 1843), the NDVI decline slows, and a slight recovery is observed. While this trend suggests a potential transition toward improved environmental conditions at higher economic levels, it does not definitively confirm the EKC hypothesis. The regression equation’s correlation coefficient (R²) is approximately 0.65, indicating that around 65% of the variation in NDVI can be explained by GDP per capita. This suggests a moderate relationship between the two variables, supporting the EKC hypothesis. A more comprehensive analysis, including additional environmental indicators and broader temporal or regional data, must conclusively validate whether the relationship follows the EKC pattern.

This study evaluates the impact of economic growth on environmental health across different regions of Bangladesh using NDVI slopes, GDP trends, and correlation analysis. Findings indicate Dhaka and Chattogram exhibit sustainable growth, balancing economic expansion and ecological stability. In contrast, regions like Rajshahi and Mymensingh show declining NDVI, suggesting economic growth may occur at the expense of vegetation health. The results emphasize the need for integrated policies that align poverty reduction with environmental sustainability. A balanced approach is essential to ensure long-term development, where economic progress does not lead to ecological degradation but fosters resilience and sustainability.

The NDVI slopes, GDP slopes, Pearson correlation coefficients, and development patterns for various regions in Bangladesh are summarized in Table 4 for three periods: 2014–2017, 2018–2020, and 2021–2023. In Dhaka, the NDVI slope remains stable, ranging from 0.003 to 0.004, which indicates consistent positive vegetation health. Meanwhile, the GDP slope increases from 1,400 to 1,600, before slightly decreasing to 1,500. The Pearson correlation coefficient reaches a peak of 0.94 during the 2018–2020 period, reflecting a strong relationship between economic growth and NDVI values. This leads to a classification of “A” for integrated sustainable development. Chattogram exhibits a similar trend, with NDVI slopes ranging from 0.002 to 0.0022 and GDP slopes increasing from 1,150 to 1,250, before dropping to 1,200. The Pearson correlation coefficients remain high, ranging from 0.83 to 0.88, resulting in a development pattern “A” classification.

In Khulna, the NDVI slope is relatively low, slightly declining from 0.0015 to 0.0018. In contrast, the GDP slope increases from 900 to 1,100, indicating economic growth in the area. The Pearson correlation coefficient ranges from 0.75 to 0.8, suggesting a weaker association than Dhaka and Chattogram. Nonetheless, Khulna maintains a development pattern classified as “A”. In Sylhet, the NDVI slope is decreasing, dropping to 0.0008 between 2018 and 2020 before experiencing a slight recovery. The GDP slope fluctuates around 950 during this time. The Pearson correlation in Sylhet remains moderate, ranging from 0.6 to 0.7. This indicates a moderate relationship between ecological health and economic growth, with Sylhet classified as development pattern “B” suggesting a stronger emphasis on economic growth over ecological health.

In Rajshahi, the NDVI slope shows a negative trend, indicating a decline in vegetation health, with values ranging from −0.0004 to −0.0006. Meanwhile, the GDP slope remains stable, fluctuating between 750 and 850. However, the weak Pearson correlation suggests a poor association between economic growth and NDVI, classifying Rajshahi under development pattern “B”. Similarly, Barishal also displays a negative NDVI slope, worsening from −0.0006 to −0.001, while the GDP slopes range from 700 to 800. The weak Pearson correlation, which varies from 0.42 to 0.4, indicates a negative relationship between economic growth and vegetation health, thus maintaining development pattern “B”.

The trends in Rangpur present a mixed picture. There is a slight positive NDVI slope, decreasing from 0.0015 to 0.0009, and a slightly increasing GDP slope, rising from 600 to 700. However, the Pearson correlation decreases from 0.7 to 0.5, indicating a weakening relationship over time. This region is classified as development pattern “C”, reflecting eco-environmental dominance. In contrast, Mymensingh shows concerning trends, with a negative NDVI slope that ranges from −0.0008 to −0.0012 and low GDP slopes between 500 and 700. The Pearson correlation here is weak, ranging from 0.35 to 0.25, which suggests a poor relationship between economic growth and NDVI. Consequently, it is classified as development pattern “D”, indicating deteriorative development. Overall, the data highlights the varying relationships between economic growth and ecological health across different regions. This underscores the need for integrated and sustainable development policies, particularly in areas experiencing negative trends.

Analyzing the relationship between NDVI values and economic growth rates in Bangladesh provides important insights into the effectiveness of poverty alleviation policies, especially in ecologically vulnerable areas. Regions with high poverty levels, often in remote or disaster-prone areas, show a strong connection between economic development and environmental health. Understanding these dynamics is key to shaping effective poverty alleviation strategies. The study found that regions with a balanced approach to economic growth and environmental protection, classified as integrated sustainable development (Pattern A), showed positive outcomes. Economic growth efforts in Dhaka and Chattogram were closely linked with environmental conservation, leading to sustainable livelihoods. The consistent NDVI trends in these regions highlight the success of policies prioritizing economic development and ecological preservation.

On the other hand, regions focused mainly on economic growth (Pattern B) faced challenges. In divisions like Mymensingh and Barishal, despite the rise in GDP, the negative trends in NDVI suggested that economic activities may have contributed to environmental degradation. This emphasizes the need for policies that promote economic growth and safeguard the environment for long-term success in poverty alleviation. Regions following ecologically driven development (Pattern C) and deteriorative development (Pattern D) showed troubling trends. In these areas, poverty alleviation efforts failed to improve living conditions and, in some cases, worsened environmental issues.

This highlights the urgent need to rethink and adjust policies in these regions. Immediate actions are needed to align economic growth with environmental sustainability to prevent further damage. This study underscores the importance of considering both economic and environmental factors when designing poverty alleviation strategies in Bangladesh. A balanced approach integrating economic growth with ecological preservation is essential for creating sustainable, long-lasting improvements in people’s lives, especially in vulnerable regions. Future policies must consider each region’s unique challenges to ensure that people and the environment thrive together.

This research highlights the complex relationship between NDVI values and economic growth rates in Bangladesh, emphasizing the crucial role of ecological health in shaping effective poverty alleviation policies. The findings indicate that regions exhibiting integrated sustainable development patterns experience economic growth while maintaining or improving their environmental health. In contrast, areas that focus primarily on economic development often suffer declines in vegetation health, suggesting a harmful trade-off that could undermine long-term poverty alleviation efforts. The analysis reveals that regions identified with ecologically driven development and those with deteriorating development patterns require immediate policy adjustments. These regions illustrate the consequences of neglecting environmental considerations in economic planning, underscoring the urgency of integrating ecological sustainability into poverty alleviation strategies. Ultimately, this research advocates for a holistic approach to development that aligns economic growth with environmental protection. By promoting policies recognizing the interconnectedness of economic and ecological systems, Bangladesh can pave the way for resilient and sustainable development, ensuring that poverty alleviation efforts improve living standards for vulnerable populations. The findings of this study contribute to a better understanding of Bangladesh’s socio-economic dynamics and offer valuable insights for policymakers looking to optimize poverty alleviation strategies in the face of environmental challenges.

We sincerely thank our supervisors, Professor Li Gu and Zhiwen Gong, for their invaluable guidance, encouragement, and constructive feedback, which have been instrumental in shaping this research.