Introduction

Polders are among the most important forms of human heritage, encompassing various types such as settlement heritage, irrigation engineering heritage, and agricultural cultural heritage1. Through long-term human-water interaction, they have given rise to sustainable systems with significant historical and cultural value, landscape and ecological value, hydraulic engineering value, and agricultural economic value2. The polder, a land-use form created through agricultural development in shallow water areas across the world, addresses flooding, tides, and agricultural production needs by methods such as digging ditches, building dikes, and installing sluices3. The origins of traditional polders in China date back to the Spring and Autumn Period and the Warring States Period (770–221 B.C.), with their main distribution in the Yangtze River Delta, the Pearl River Delta, and the Dongting Lake area4.

For centuries, polder systems have recorded and preserved the harmonious interaction among agricultural, social, economic, and ecological elements, forming the fundamental geographical units of river network plains in southern China5. They carry a rich legacy of both tangible and intangible cultural heritage, serving as valuable public cultural resources that warrant both preservation and development. Traditional polders represent more than agricultural infrastructure; they are multifunctional heritage systems that embody ecological regulation, cultural continuity, spatial wisdom, and social organization. Recognizing their composite value is essential for framing effective sustainability and conservation strategies. However, climate change and the rapid expansion of urban areas have significantly altered the landscape patterns of traditional polders, intensifying the conflict between land use and ecological risks6. The original water systems and traditional architectural structures are gradually disappearing, posing a severe threat to the protection and development of traditional polders heritage7. Sustainable development is often pursued through multidimensional and integrated approaches8. Effective conservation measures can promote the sustainability of economic-social-ecological systems while preserving cultural heritage, which is of paramount importance for the development of human society9.

In the 1950s, research on traditional polders was just beginning, with most studies focusing on basic issues such as the formation background, historical evolution, and water conservancy infrastructure, approached from the perspectives of history, hydrology, and related fields10,11. Since the 1980s, the development of agricultural economics and sociology has shifted the focus of research to the technological advancements and changes in social relations associated with traditional polders4,12,13. In the 21st century, the intersection of traditional polders with fields such as agriculture, ecology, economics, and landscape studies has given rise to new research methods and perspectives. Nijhuis advocates for using spatial planning methods to protect Dutch polders, integrating agricultural, urban, and landscape elements to preserve the cultural and ecological functions of these heritage landscapes in the long term14.Boer proposed the European Polder Program, aiming to develop a network of European polder landscapes with cultural and natural value by establishing regional innovation sites. The program seeks to benefit the regional economy while enhancing the accessibility and visibility of polder heritage15. Winaktoe applied the Sustainable Urban Polder Model (SUPM) to the Jakarta polders, aiming to fill the gaps in the integration of technology, finance, and institutional aspects in Jakarta’s flood control research. He emphasized that establishing a polder committee is crucial for enhancing the sustainability and functionality of these polders16. Various studies show that the functional value of polders has been preliminarily recognized and explored on a global scale.

In recent years, research on flood control, planning and construction, ecological environmental changes, and socio-economic aspects of polders has received increasing attention. Sun et al. analyzed the regional characteristics of the traditional polder settlement system in the Xitiaoxi River Basin from a morphological perspective, proposing sustainable ecological strategies for flood and disaster management. They also used ArcGIS to reveal the evolution of the polder landscape shaped by natural and hydrological factors17. Li et al. reconstructed the distribution and changes of polders in the Dongting Lake Basin from 1368 to 1980 using remote sensing to extract polder zoning and historical maps, and explored the wetland loss caused by polder construction18. Santoro analyzed the land use changes in the Mulberry Fish Pond system in Huzhou, China, caused by factors such as urban expansion and the construction of photovoltaic systems. He proposed that strengthening local planning tools for the protection of this cultural heritage is an urgent need19. E.C. Ellis et al. examined the ecological basis for the sustainable development of polder agriculture in the Taihu Lake region of China within a historical context20. Yi-Wen Wang et al. have described it as a sustainable landscape, emphasizing its importance in dynamic socio-ecosystems1. Although these large-scale studies indicate that the importance of traditional polder heritage is increasingly recognized, and various analytical methods have been used to study its spatial evolution, land use patterns and ecological security, the deeper driving factors behind these changes remain to be explored. While traditional polders have historically demonstrated high productivity20, questions about whether the system’s natural supply capacity can continue to meet the demands of modern society, and whether it can maintain a positive interaction between economic, social, and ecological aspects, still lack sufficient theoretical and empirical support.

With the advancement of society, the public’s demand for and utilization of cultural heritage continue to evolve. A key challenge that urgently requires resolution is how to ensure the long-term preservation of cultural heritage while meeting the needs of contemporary society. In 2015, the United Nations adopted the 2030 Agenda for Sustainable Development, which recognized the long-term role of cultural heritage in urban development, and established 17 Sustainable Development Goals (SDGs)21, and the New Urban Agenda, adopted at the United Nations Conference on Housing and Sustainable Urban Development (Habitat III) in 2016, identified both tangible and intangible heritage as essential elements of sustainable urban development22. Nocca noted that, although the role of cultural heritage in sustainable development has been widely discussed at international forums, its practical implementation remains limited, often confined to theoretical discourse. To address this gap, she proposed using empirical research to demonstrate the contributions of heritage across economic, social, and environmental dimensions, and constructed a matrix of sustainability indicators23. Schmitz and Herrero-Jáuregui focused on socio-ecological sustainability in the protection of cultural landscapes, highlighting the growing need for innovative theories and methodologies to safeguard both cultural and natural heritage9.

The Ecological Footprint (EF) is an indicator that reflects the demand of humanity for biological resources, used to assess the extent of resource consumption resulting from human activities24. Ecological carrying capacity (EC) refers to the ability of a region to continuously provide resources and absorb waste, reflecting the potential for the sustainable supply of natural resources in that region24. In the process of achieving sustainable development for human society, maintaining the stock of natural resources and ensuring their sustainable use are key, and the dynamic balance between the two is crucial8,25. Since the concept of EF was introduced by economist Rees26 in 1992, the model has been widely used in fields such as ecological security assessment and sustainable utilization evaluation, covering spatial scales from global to national and regional levels27,28,29,30. In recent years, some scholars have applied the EF model to individual studies of biologically productive lands (such as arable land, forest land, grassland, water areas, built-up areas, and energy land)25,31,32.Traditional polders not only bear historical and cultural value but also face pressures from population growth, economic development, and technological advancements in modern society3. The EF model can integrate these multidimensional driving factors into quantitative analysis, helping assess the carrying capacity of traditional polders and the challenges they face.

In response to the Sustainable Development Goals33 and to promote the long-term development of cultural heritage, this study adopts an interdisciplinary approach by applying the EF model to assess the sustainability of traditional polder heritage. Conventional EF models typically assume that land serves only productive functions. However, in practice, agricultural production and daily rural life generate significant carbon emissions. Meanwhile, traditional polder practices–such as rice-fish co-culture and mulberry-dyke-fishpond systems–exhibit considerable carbon sequestration potential. To address the dimensional limitations of conventional EF models, this study incorporates carbon footprint accounting, thereby capturing both emission and absorption processes within polder ecosystems.

Focusing on Nanxun District in Huzhou, Zhejiang Province, China, this study constructs a dynamic model of supply-demand balance to better reflect the specific characteristics of this complex socio-ecological system. Furthermore, it employs the Stochastic Impacts by Regression on Population, Affluence, and Technology (STIRPAT) model and Partial Least Squares (PLS) regression to identify the effects of population, affluence, and technology, adapted to the unique geographical context of Nanxun. The goal is to reveal how these factors influence the EF of polders, providing both theoretical insights and empirical evidence to support the positive interplay between traditional polders and their ecological, economic, and social dimensions, while offering a novel framework for integrating heritage conservation with sustainability assessment.

Methods

Study area

Huzhou, located in the southern part of Taihu Lake in Zhejiang Province, China, is abundant in water resources, including rivers, lakes, and wetlands, and has a subtropical monsoon climate (Fig. 1). Huzhou is both the birthplace and one of the best-preserved areas of the traditional polders (Fig. 2) in the Taihu Lake Basin. The polder systems here comprise three main components: agriculture, water management, and settlement17. The agricultural system comprises rice paddies, drylands, and fish ponds. Rice paddies are predominantly used for rice cultivation, while drylands support crops such as vegetables, wheat, and maize. Fish ponds are primarily dedicated to the farming of fish and crabs. The water management system includes embankments, river networks, sluices, and shallow lakes, and encompasses the Lougang irrigation and drainage system in the Taihu Basin, a World Heritage Irrigation Structure. Lougang refers to the narrow, densely distributed rivers built to drain water into Taihu Lake7. Originally, there were 74 Lougang in Huzhou, of which 66 remain well-preserved today. Only 21 of these directly flow into Taihu Lake7. The silt accumulated in these rivers is used for the cultivation of mulberry trees, thus forming the Mulberry Fish Ponds, also known as Mulberry Polder, which represent an integrated ecological cycle and highly efficient agriculture. This system is recognized as a Globally Important Agricultural Heritage System19.

Fig. 1
figure 1

Location of study area.

Full size image
Fig. 2
figure 2

The traditional polder and settlement.

Full size image

According to statistics, there were 1676 polders in Huzhou in 1990, covering an area of 1394.2 square kilometers34.In addition to the reduction in number, land use in the traditional polders has also undergone significant changes. The area devoted to fish ponds has steadily increased due to the substantial economic benefits of aquaculture, while the areas of mulberry gardens and rice paddies have notably declined. From 2005 to the present, the area used for crop cultivation has decreased by approximately 43.4%, while the area dedicated to freshwater aquaculture has increased by about 29.3%. Currently, Nanxun District covers an area of approximately 702.26 km2. As of March 2024, the Mulberry Fish Pond system in Nanxun retains 40 km2 of mulberry orchards and 100 km2 of fish ponds, making it the region with the highest concentration, the largest area, and the most well-preserved Mulberry Fish Ponds globally (Fig. 2. However, with the acceleration of urbanization, some Mulberry Fish Ponds have been converted into construction land for real estate development35, leading to a decline in the agricultural and ecological functions originally provided by the traditional polder systems6. As a Globally Important Agricultural Heritage System recognized by the Food and Agriculture Organization of the United Nations, selecting the traditional polder heritage of Nanxun District as a research case offers a valuable example for promoting the sustainable development of cultural heritage worldwide.

Sources of data

The indicators presented in Table 1 are categorized into four main systems: (1) biological indicators, (2) socioeconomic indicators, (3) yield and equivalence indicators, and (4) carbon footprint indicators. The core dataset primarily comes from official statistical yearbooks and academic literature. Specifically, biological and socioeconomic indicators are derived from the Huzhou Statistical Yearbook and Zhejiang Statistical Yearbook (2005–2022 editions). Yield and equivalence, and carbon footprint indicators follow standardized values from Zhang et al. (2009) based on sub-national hectare calculations36. Carbon footprint data–including emission and absorption coefficients–are sourced from peer-reviewed studies on farmland ecosystems in Yunnan and Dezhou, ensuring consistency with prior research methodologies37,38.

Table 1 Classification of indicators and data sources
Full size table

Improved EF and EC model

Demand model: EF of traditional polders

This study introduces the measurement criterion of provincial hectare, which represents the average biological productivity of one hectare of land at the provincial level, providing a more precise assessment of the sustainable development of the traditional polder heritage in Nanxun District39. The original EF model measures the consumption of natural resources by human activities, assessing whether human demand exceeds the Earth’s sustainable supply by converting resource use into corresponding land areas28,30. However, the model assumes that each type of land serves only biological production functions, focusing primarily on the supply and consumption of primary products. Traditional polders, beyond providing resources for human living and production, also play a crucial role in maintaining the regional ecological environment40. Considering the unique connotations and functions of traditional polders, this study categorizes the EF into ecological productive footprint (EP) and ecological output footprint (EO)41. The EP of traditional polders refers to the human consumption and utilization of their natural resources26. The specific calculation formulas are as follows:

$$EF=EP+EO$$
(1)
$$ep=\frac{EP}{N}=\mathop{\sum }\limits_{i=1}^{n}\frac{{A}_{i}}{N}=\left(\mathop{\sum }\limits_{i=1}^{n}{r}_{j}\times \frac{{P}_{i}}{E{P}_{i}}\right)/N$$
(2)

In the formula above: ep is EP per capita (ha); Ai is biologically productive area corresponding to the i-th consumption item (ha); N is the total population; Pi is annual production of the i-th consumption item (kg); EPi is the average productivity of the i-th consumption item within the province (kg/ha); rj is the equivalence factor for the j-th type of biologically productive land. The equivalence factor of each biologically productive land is shown in Table 1. n is the number of crop types, set to n = 9 in this study.

Based on the production characteristics of traditional polders in Nanxun District, nine crops–rice, wheat, maize, soybean, cotton, rapeseed, potato, peanut, and vegetables–along with freshwater aquaculture products, were selected as the primary consumption items42.

The EO refers to the impact of waste generated by human production activities on traditional polders37,43. The model is based on the carbon footprint accounting method for agricultural ecosystems37, with the specific calculation formula as follows:

$$eo=\frac{EO}{N}=\frac{E/\left(\frac{W}{S}\right)}{N}$$
(3)

In the formula above: eo is carbon footprint per capita (ha); E is carbon emissions (kg); PC is carbon sequestration capacity (kg); W is carbon absorption (kg); S is the area of traditional polders (ha).

Traditional polders heavily utilize fertilizers, pesticides, and other production tools in agricultural activities, leading to substantial carbon emissions. The calculation formula is as follows:

$$E={E}_{f}+{E}_{p}+{E}_{m}+{E}_{e}+{E}_{i}$$
(4)

In the formula above: E is the total carbon emissions (kg); Ef, Ep, Em, Ee, Ei are the carbon emissions (kg) resulting from fertilizer application, pesticide application, agricultural plastic film production and usage, agricultural machinery usage, and agricultural irrigation, respectively.The calculation formulas for each component are as follows:

$${E}_{f}={G}_{f}\times A$$
(5)
$${E}_{p}={G}_{p}\times B$$
(6)
$${E}_{m}={G}_{m}\times C$$
(7)
$${E}_{e}=({A}_{e}\times D)+({W}_{e}\times F)$$
(8)
$${E}_{i}={A}_{i}\times G$$
(9)

In the formula above: Gf is the FU (kg); Gp is the pesticide usage (kg); Gm is the agricultural plastic film usage (kg); Ae is the crop planting area (ha); We is the total agricultural machinery power (kW); Ai is irrigation area (ha). A, B, C, D, F and G represent the carbon emission coefficients for fertilizer, pesticide, agricultural plastic film, crops, agricultural machinery, and irrigation, respectively. The determined values are as follows44:

$$A=0.89\,\,{\text{kg/kg}}\,;\quad B=4.93\,\,{\text{kg/kg}}\,;\quad C=5.18\,\,{\text{kg/kg}}\,;$$
$$D=16.47\,\,{\text{kg/ha}}\,;\quad F=0.18\,\,{\text{kg/kW}}\,;\quad G=266.48\,\,{\text{kg/ha}}\,.$$

Quantitative calculation of carbon absorption for different crops in polders requires considering crop yield, economic coefficient, root-to-shoot ratio, and carbon content. The calculation formula is as follows:

$$W=\mathop{\sum }\limits_{i=1}^{n}{W}_{i}=\mathop{\sum }\limits_{i=1}^{n}\left[{C}_{i}\times {K}_{i}\times (1-{V}_{i})\times (1+{R}_{i})\right]/{H}_{i}$$
(10)
$$D=\frac{W}{A}=\mathop{\sum }\limits_{i=1}^{n}\frac{{W}_{i}}{A}$$
(11)

In the formula above:W is the total carbon absorption (kg); Wi is carbon storage of the i-th crop (kg); Ci is carbon content (%); Ki is yield of the i-th crop (kg); Vi is moisture coefficient of the i-th crop fruit (%); Ri is root-to-shoot ratio coefficient of the i-th crop; Hi is economic coefficient of the i-th crop (ratio of economic yield to biological yield); D is the average carbon density of the traditional polder ecosystem (kg/ha); A is crop planting area (ha). The values of H, C, R, and V for major crops in China are shown in the Table 237,38.

Table 2 Estimation Parameters for Carbon Storage of Different Crops
Full size table

Supply model: EC of traditional polders

EC evaluates the PS or economic activity intensity that a specific region or ecosystem can sustainably support without jeopardizing its long-term health45. It represents the extent to which polders can sustain human activities. The calculation formula is as follows:

$$ec=\frac{EC}{N}=\left[\mathop{\sum }\limits_{i=1}^{n}{S}_{i}\times {y}_{i}\times {r}_{i}\times (1-x)\right]/N$$
(12)

In the formula above: Si is area of the i-th consumption item (ha); yi is yield factor of the i-th consumption item; x is the biodiversity conservation area (BCA). The yield factor is calculated on the provincial hectare scale, defined as the ratio of the crop yield per unit area in Nanxun District to that in Zhejiang Province. According to the report Our Common Future by the World Commission on Environment and Development, 12% of the land area should be allocated for biodiversity conservation in the calculations46. As this study focuses exclusively on traditional polders, the annual deduction for BCA was determined using the area-weighted method to ensure scientific rigor25.

Evaluation of the sustainable use of traditional polders

Ecological pressure index

The Ecological Pressure Index (EPI) reflects the relationship between the supply of ecological resources from traditional polders and human demand30. Higher values indicate a greater intensity of ecological pressure on the environment. The specific grading criteria for ecological security are detailed in Table 3 and the formula is shown as follows29:

$$\,{\text{EPI}}\,=\frac{ef}{ec}$$
(13)
Table 3 Classification standard of ecological security
Full size table

Ecological surplus and deficit

Ecological surplus (ES) occurs when the consumption of polder resources by human activities is less than the polders productive capacity, indicating regional sustainability30. In contrast, ecological deficit (ED) arises when human development surpasses the EC of the polders, reflecting an unsustainable state. The formula is shown as follows:

$$ED=EC-EF$$
(14)

Ecological sustainability index

Building on the analysis of ES and ED, this study introduces the Ecological Sustainability Index (ESI) to evaluate the sustainability of traditional polders more comprehensively47. The ESI measures the extent to which EC meets EF, with higher values indicating stronger sustainability. If 0 < ESI < 0.5, the EF exceeds the EC, indicating an unsustainable state. When ESI = 0.5, it signifies that the EF and EC are in balance. Finally, if 0.5 < ESI < 1, it indicates that the EF is below the EC, reflecting a sustainable state. The formula is shown as follows:

$$\,{\text{ESI}}\,=\frac{EC}{EF+EC}$$
(15)

The STIRPAT modelling approach

STIRPAT is an extended statistical analysis model used to study the relationship between environmental impact and population, affluence, and technology. It was proposed by Dietz48 and others as an improvement to the IPAT49 model, and its formulation is as follows:

$$I=a{P}^{b}{A}^{c}{T}^{d}\epsilon$$
(16)

In the formula above: a is the model coefficient; b, c, and d denote the driving force indices for population, affluence, and technology; ϵ is model error term, reflecting the contribution of uncertainties and other factors to environmental impact. By taking the logarithm of both sides of the formula:

$$\ln I=\ln a+b\ln P+c\ln A+d\ln T+\ln \epsilon$$
(17)

This study uses the EF to characterize the impact of human activities on traditional polders, comprehensively considering the interaction between human socio-economic development and traditional polders, and expanding the three factors in the model50,51. First, PS is a fundamental factor influencing the expansion of EF28. Second, urbanization is the core component and primary driver of regional economic development; therefore, the affluence indicators are represented by the UR and per capita GDP30. Lastly, The technological factor represents all elements that are not related to population and affluence52. Therefore, based on the characteristics of agriculture and water resources in traditional polder heritage, this study selected four indicators: AM, FU, IMP, and FMP53. Thus, the extended STIRPAT model for the influencing factors of the EF of polders in Nanxun District is formulated as follows:

$$\begin{array}{rcl}\ln EF&=&\ln a+b\ln (PS)+c\ln (UR)+d\ln (GDP)+e\ln (AM)\\ &&+f\ln (FU)+g\ln (IMP)+h\ln (FMP)+\ln \epsilon \end{array}$$
(18)

To address the issue of multicollinearity in traditional linear regression models and to extract stable regression relationships in high-dimensional data, this study employs the PLS regression method to analyze the STIRPAT model54.

Results

Evaluation results of EF and EC

From 2005 to 2022, the total EF of polders in Nanxun District showed an overall declining trend with fluctuations, decreasing from 139,889.638 ha to 71,056.866 ha, a reduction of 49.2% (Fig. 3, Table 4). Despite the overall decline, there were periods of temporary increases, such as during 2005-2006, 2008-2009, and 2020-2021. Notably, a sharp decline occurred between 2013 and 2014, when the EF dropped from 132,094.459 ha to 93,164.085 ha. The lowest point was recorded in 2018, with an EF of 70,776.842 ha. Changes in the EF are closely linked to fluctuations in crop planting area and yield. The significant decline in the EF of traditional polders between 2013 and 2014 was primarily driven by increased production costs. For instance, data from the Compilation of Cost-Benefit Data of Major Agricultural Products in China reveal that the total cost of rice production per hectare in Zhejiang Province reached 18,292.20 yuan in 2014, an increase of 10,489.65 yuan compared to 2005. The substantial rise in production costs significantly impacted the profitability of rice farming. Furthermore, the lowest crop planting area, recorded at 24,319 ha in 2018, contributed to the EF reaching its minimum value.

Fig. 3
figure 3

Changes of EF, EC, ef, and ec in study area from 2005 to 2022.

Full size image
Table 4 Changes in Indicators within the EF Model
Full size table

Over the past 18 years, the EP of traditional polders in the study area has consistently been much larger than the EO, indicating significant occupation of the polders due to human consumption of agricultural and freshwater aquaculture products (Fig. 4, Table 4). The EO initially declined from 5042.864 ha, rebounded to 5026.736 ha in 2010, reached a minimum of 3399.932 ha in 2018, and then increased again to 3998.445 ha by 2022. The changes of EO are primarily driven by the use of agricultural fertilizers and pesticides; the increase in both factors has contributed to a heavier ecological burden on traditional polders (Fig. 5).

Fig. 4
figure 4

Changes of EP and EO in study area from 2005 to 2022.

Full size image
Fig. 5
figure 5

Agricultural fertilizer and pesticide usage from 2005 to 2022.

Full size image

From 2005 to 2022, the overall trend of the total EC and the EF of traditional polders in Nanxun District was closely aligned, exhibiting a fluctuating downward trend (Fig. 3, Table 4). However, the EC consistently remained below the EF, with the difference between the two narrowing over time. The EC declined from an initial 130,804.603 ha to 73,488.949 ha, representing a 43.8% decrease. There was a temporary recovery between 2008 and 2009, but a sharp decline occurred between 2013 and 2014, dropping from 115,403.607 ha to 82,981.371 ha, a 28.1% reduction. The primary reason for this was a significant decrease in the area cultivated with crops. According to the 2023 Huzhou City Environmental Status Report, from 2005 to 2015, Huzhou City experienced significant acid rain pollution, with the annual average precipitation pH values ranging between 4.29 and 5.34, and the annual average frequency of acid rain between 70.8% and 97.9%. Notably, from 2013 to 2015, the acid rain occurrence rate in Huzhou City increased significantly (Fig. 6). Acid rain led to soil acidification and a substantial decline in crop yields–particularly staple grains–which resulted in a sharp drop in the EC of traditional polders during this period (Fig. 3). After 2015, the acid rain occurrence rate showed an overall declining trend, reaching its lowest point in 2022. Following several years of recovery from soil acidification, the EC reached its minimum in 2019 (Fig. 3) and began a gradual upward trend thereafter.

Fig. 6
figure 6

Trends of mean pH and acid rain rate of precipitation in Huzhou city from 2005 to 2022.

Full size image

Overall, the EF and EC in Nanxun District are influenced by three main factors:

  1. 1.

    With the acceleration of urbanization and industrialization in Nanxun District, land use planning has changed, leading to the requisition of some polders for industrial and infrastructure development.

  2. 2.

    Climatic disasters such as floods and acid rain have resulted in a reduction in the area of traditional polders or a decrease in crop yields.

  3. 3.

    As the economic benefits of sericulture and traditional grain cultivation are lower than those of freshwater aquaculture, there is a shift towards more economically efficient crops and an increase in aquaculture, particularly fish farming (Fig. 7). However, the equivalence factor of fishponds is lower than that of cultivated land, which also contributes to a decrease in the total EC.

    Fig. 7
    figure 7

    Changes of crop cultivation area and freshwater aquaculture area in study region from 2005 to 2022.

    Full size image

Evaluation results of sustainable utilization of traditional polders

Figure 8 and Table 5 indicate that the ESI for traditional polders in Nanxun District predominantly hovered at low levels from 2005 to 2017. Although a distinct peak of 0.485 was observed in 2018, it declined to 0.469 by 2019. The ESI reached a pivotal turning point in 2021, ascending to 0.508 in 2022. Meanwhile, the EPI consistently ranged between 1.063 and 1.150 throughout 2005 to 2021. According to the EPI grading evaluation standard table (Table 3), the ecological pressure on traditional polders has consistently been in a critical overload state. Similarly, traditional polders exhibited an ED from 2005 to 2021. These three indicators collectively suggest that from 2005 to 2021, traditional polders were in an unsustainable state. In 2022, although the EPI and ESI only slightly exceeded the critical values by 0.033 and 0.008, respectively, and the ES value was also small, traditional polders have not yet achieved a fully sustainable status. However, since 2005, ED and ESI have shown a fluctuating upward trend, and the EPI has gradually declined, which sufficiently indicates that the sustainability of traditional polders has improved.

Fig. 8
figure 8

Evolution of ED, ES, ESI, and EPI in study area from 2005 to 2022.

Full size image
Table 5 Changes of ecological security in study area from 2005 to 2022
Full size table

Identification of driving factors for the EF of traditional polders

Based on the results of Pearson correlation analysis (Table 6), it is evident that multicollinearity exists among the factors affecting the EF in Nanxun District. Therefore, PLS is used for further analysis.

Table 6 Correlation matrix of independent variables
Full size table

By extracting a single latent factor, the model achieves a goodness of fit of 85.9% (Table 7), indicating a high level of model fitting. Variable Importance in Projection (VIP) analysis elucidates the relative importance of each driver on the EF (Fig. 9). In Nanxun District, FU has the highest VIP value among the driving factors, at 1.318, indicating its greatest contribution to explaining EF. Following this, GDP and UR both exceed the empirical threshold (VIP > 1), with weights of 1.180 and 1.097, respectively, highlighting the significant role of economic activity levels in influencing EF. The VIP values for AM and PS are close to the threshold of 1. Although they do not meet the significance standard, they still provide some explanatory power to the model. The VIP value for PS (0.825) is comparatively low, indicating that PS has a relatively small contribution to the model. From 2005 to 2022, although there were fluctuations, the overall change in PS was minor, generally showing a slight downward trend (Fig. 10). This relatively stable population trend accounts for the low score in the VIP analysis. The VIP values for IMP and FMP are both less than 1, indicating their minor contribution to the model. However, further analysis in conjunction with the regression coefficients is needed.

$$\begin{array}{ll}EF\,=\,-126.821\times {(PS)}^{8.979}\times {(UR)}^{-0.289}\times {(GDP)}^{-0.123}\times {(AM)}^{1.548}\\\qquad\qquad \,\times \,{(FU)}^{0.275}\times {(IMP)}^{-0.105}\times {(FMP)}^{-0.1}\end{array}$$
(19)
Table 7 Proportion of Explained Variance
Full size table
Fig. 9
figure 9

The VIP analysis of affecting factors of EF.

Full size image
Fig. 10
figure 10

Changes of population in study area from 2005 to 2022.

Full size image

Based on the PLS regression analysis results, the STIRPAT model for the traditional polder heritage is restored. The formula shows that PS, AM, and FU are positively correlated with EF in the study area, with impact indices of 8.979, 1.548, and 0.275, respectively. This indicates that population growth, increased AM, and the rising use of fertilizers are the main driving factors behind the expansion of EF, with PS having the most significant impact. In contrast, the growth of UR, GDP, IMP, and FMP acts as negative driving forces for EF, with indices of −0.289, −0.123, −0.105, and −0.1, respectively.

Discussion

The primary aim of this study is to evaluate the sustainability of traditional polder heritage utilization and its influencing factors. Through an extended EF model incorporating carbon footprint analysis and the STIRPAT framework, we quantitatively assessed the relationship between natural supply and human demand, as well as the external drivers of sustainability in Nanxun District. The main conclusions are as follows:

From 2005 to 2021, the traditional polder systems exhibited an unsustainable trajectory, marked by long-term ecological deficits. A minor ES appeared in 2022; however, the ESI exceeded the threshold by only 0.008, suggesting that true sustainability has not yet been achieved. Both the ef and ec showed a fluctuating downward trend, and the ED gradually narrowed. The EPI rating dropped from “Moderately Risky” to “Moderately Safe,” reflecting some improvement in ecological conditions. Land-use planning, climatic shocks, and the economic return of polder crops were identified as key direct factors influencing EF and EC. Among the STIRPAT indicators, PS, AM, and FU positively contributed to EF, while UR, GDP, IMP, and FMP played mitigating roles.

Modernization has disrupted the traditional ecological balance, and the natural supply capacity of polders can no longer meet growing demands. Land-use change emerged as the most significant pathway through which PS, AM, and FU indirectly increase EF and reduce EC. These findings support previous arguments that land-use planning is central to balancing ecological functions and heritage conservation55. Proper regulation can reduce ecological stress and protect the structural integrity of polder landscapes56.

Demographic pressure remains a core concern. Although Nanxun has experienced relatively stable population levels, mobility associated with urbanization and tourism inflates the effective population footprint41. A zoning-based ecological regulation and early-warning mechanism integrating real-time EF and mobility monitoring should be developed to prevent ecological overshoot in heritage zones.

Affluence, represented by GDP and UR, has a statistically negative correlation with EF. As observed in other regions37,57, increasing economic capacity facilitates investments in water infrastructure, environmental governance, and green innovation, all of which improve system resilience. However, in agricultural regions like Nanxun, affluence-driven land conversion may offset these benefits unless paired with regulatory land-use frameworks53. Promoting eco-agritourism could create a feedback loop where income from cultural tourism directly supports polder maintenance and sustainability.

Technological factors had mixed effects. AM and FU significantly raised EF, particularly where fertilizer use boosts yields at the cost of soil health and long-term productivity20,27,37,58. In contrast, IMP and FMP were associated with reduced ecological pressure, highlighting the sustainability benefits of water-efficient infrastructure in irrigation and aquaculture.

To translate these findings into actionable strategies, we propose the following three-tiered framework:

  • Ecological Management: Encourage the restoration of high-carbon sequestration agro-ecological practices such as rice-fish co-culture and mulberry-dyke fishpond systems19; Integrate rice paddy carbon sequestration capacity into local carbon trading schemes and provide subsidies for low-carbon inputs.

  • Spatial Regulation and Zoning: Establish zoning systems to separate tourist and agricultural polder zones with capacity-based thresholds; Combine EF monitoring with real-time population mobility data to trigger early-warning systems for ecological overload.

  • Economic and Institutional Incentives: Promote eco-friendly rural tourism that reinvests tourism revenue into polder conservation; Support policies that link GDP growth with environmental and heritage preservation, fostering institutional support for sustainable polder infrastructure.

Despite its contributions, this study has limitations. The sustainability of traditional polders is shaped not only by biophysical and technological drivers but also by intangible cultural, hydrological, and institutional elements. While this study used robust time-series data from 2005–2022, the lack of long-term historical datasets prevents comparative analysis with ancient polder systems. Moreover, the current model does not incorporate non-quantifiable heritage elements such as spatial rituals and local water governance traditions. Future research could develop a “digital polder monitoring" system supported by a multi-source data platform (combining remote sensing, yearbooks, and community surveys), and a dynamic ecological stress database to enable real-time sustainability assessments.