Assessing the interconnected effects of policy interventions on shrimp farming expansion and mangrove ecosystems through system dynamics

Since the early 2000s, Ca Mau has experienced a rapid boom in shrimp farming, leading to the conversion of mangroves into shrimp ponds and impeding mangrove forest conservation. Despite its negative environmental impacts, shrimp aquaculture remains vital to Ca Mau’s economy by providing employment opportunities and contributing to the province’s gross domestic product. This study constructed a system dynamic model to analyze the complex system of shrimp aquaculture and mangrove forests under two development scenarios: the Business-as-Usual (BAU) scenario and the Policy scenario. In the BAU scenario, shrimp aquaculture will continue to expand, resulting in the conversion of more mangroves into ponds and a decrease in Ca Mau’s mangrove forest to 70,349 (± 888.801) hectares in 2050. However, this expansion will bolster rural employment and the province’s economy, generating 14,250 (± 0.336) billion VND (US$ 570 million) in 2050. Conversely, in the Policy scenario, stabilizing shrimp areas at 280,000 hectares as a policy target will regulate mangrove conversion, allowing mangroves to regenerate (77,016 (± 687.155) hectares in 2050) and enhancing carbon storage (65 × 106 (± 0.58 × 106) MgC in 2050). However, challenges arise in the Policy scenario concerning potential economic stagnation, conflicts with other development priorities, and rural job losses. Officials must consider more than just the area of shrimp ponds to achieve sustainable development. Effective land use strategies should be implemented to ensure equilibrium between shrimp aquaculture and mangroves. Diversifying economic activities and promoting alternative livelihoods can mitigate the dependence on shrimp farming and offset the effects of policy interventions.


Introduction
Aquaculture, particularly shrimp farming, is one of the world's fastest-growing food production industries (Galappaththi and Berkes 2015).Shrimp production worldwide increased from 1,600 tons in 1950 to 9.4 million tons in 2022 (GLOBEFISH 2023).However, this industry's rapid expansion has resulted in environmental degradation, particularly the conversion of coastal ecosystems into shrimp ponds (Barbier and Cox 2003, Pushpam 2012, Nasar et al 2017, Naila and Salman 2018).Between 1980 and 2000, the world experienced a loss of 35% of its mangrove forests, with shrimp farming being the major driver (Valiela et al 2001).The relationship between shrimp farming development and mangrove conservation has been a long-standing concern in largescale shrimp farming nations, including Thailand, Indonesia, Taiwan, China, Malaysia, and Bangladesh.
Since 1960, Southeast Asia (SEA) governments have prioritized economic development over environmental conservation (UNEP/COBSEA 2010).Consequently, the region has experienced severe environmental degradation, such as massive forest cover and habitat loss in its coastal areas.Over the past 55 years, the SEA region has lost 50% to 80% of its mangrove forests and seagrass beds (Wilkinson et al 2006).From 2000 to 2015, the SEA islands lost 2,503 km of natural coastline while adding 3,035 km of man-made shoreline (Zhang and 2. Methodology 2.1.Methods Previous research has sought to develop various approaches for investigating how shrimp aquaculture interacts with other factors within coastal ecosystems.For example, research has explored the trade-offs between mangroves and shrimp farming in Ecuador (Parks and Bonifaz 1994), evaluated the multifaceted value of mangroves under different management scenarios in Indonesia (Jack Ruitenbeek 1994), and assessed the compromises between development and pollution within ecosystem-fisheries linkages in the US (Kahn and Kemp 1985, McConnell and Strand 1989, Swallow 1994).However, prior investigations often utilized static models to evaluate the relationship between aquaculture and the environment.To achieve sustainable development in aquaculture, it is essential to balance ecological and economic advantages (Isa et al 2021).Thus, the dynamic and interconnected nature of the environment and economic activities must be considered.Consequently, this study undertakes an assessment of the relationship between shrimp farming expansion and other factors, especially mangrove conservation, by examining the interactions among three primary dimensions: economic, social, and environmental.A thorough examination of numerous important variables is undertaken within each dimension, aiming to provide a holistic evaluation of the issue.
Given the complex dynamics of shrimp aquaculture and mangrove conservation, this study conducts analyses and predictions using System Dynamic Modeling (SDM).J W Forrester developed this system at the Massachusetts Institute of Technology (MIT) in 1950.SDM is a methodology that allows for the simultaneous evaluation of multiple modules as well as stock and flow variables within a system through a set of simultaneous difference equations (Bala et al 2016).Previous research also adopted SDM to acquire insights into the complexity and temporal dynamics of coastal ecosystems (Mavrommati et al 2013, Downing et al 2014, Nila Rekha et al 2015, Huanhuan et al 2016, You et al 2018, Isa et al 2021).The SDM model in this study was segmented into sectors encompassing mangrove areas, carbon sinks, shrimp ponds, economics, and employment.We conceptualized the feedback diagram and the stock-flow diagram using Vensim (Ventana Systems 2023).Subsequently, we developed equations within the model and performed analyses in R using the packages 'deSolve' (Soetaert et al 2010) and 'ggplot2' (Wickham et al 2016).
Following the approach of Downing et al (2014), the system dynamics (SD) model in this study underwent several steps in its construction.In step 1, we formulated a feedback diagram delineating the essential components of the system, their interrelationships, and the associated directional signs.To mitigate potential biases in model development, we sought insights from experts in diverse domains, including fisheries and aquaculture sciences, environmental economics, and social sciences.Based on these consultations, we integrated the carbon subsector into the model, recognizing mangrove forests' remarkable carbon sequestration potential in Ca Mau (Tue et al 2014, Nam et al 2016).In step 2, we conducted the loop analysis to identify feedback loops that explain the dynamic interactions and chain reactions between key elements in the model.Moving to step 3, a stock-flow diagram was constructed based on the feedback model and loop analysis.We also defined each variable in the model and identified the mathematical relationships between these variables.Subsequently, in step 4, several tests were conducted to verify the constructed model, following the guide from Duggan (2016).Finally, in step 5, the verified model was used for simulation and analysis.

Mangrove sector
The mangrove forest was modeled as a renewable resource to capture changes in mangrove area over time in correlation with the development of shrimp ponds and land availability (figure 1).These feedback loops allow us to better understand the effects of mangrove restoration and degradation on the overall system dynamics.Furthermore, we assumed that a certain part of abandoned shrimp ponds will eventually revert to mangrove habitat (Arquitt et al 2005).In fact, Landsat imagery from 1999 to 2022 showed a shift in land use from shrimp ponds to mangroves (Clark Labs 2022).Policymakers and authorities often fail to recognize the reliance of the shrimp industry on ecological services.As a result, shrimp ponds can grow beyond ecological carrying capacity, especially within mangrove areas, with substantial environmental repercussions and potential system collapse (Arquitt et al 2005).
The Effect of land availability variable captured the competition for land between mangroves and ponds within the system.It estimated the extent to which land availability influences the growth or decline of mangrove areas over time.As the pond area expands, it reduces the available land for mangroves, potentially constraining their growth.Conversely, if the pond area decreases, it frees up more land for mangrove expansion.This feedback mechanism allows researchers and policymakers to explore the interplay between mangrove and pond areas and understand how changes in land availability can impact the growth or decline of mangroves within the system.Coastal areas in India and Mexico also experienced competition between anthropogenic activities and mangrove areas (Vaz 2014, Osland et al 2018).

Carbon sector
Given the high carbon sequestration capacities of the mangrove forests in Ca Mau (Nam et al 2016, Nguyen et al 2023), it is crucial to analyze how the carbon stored in these ecosystems changes in tandem with shifts in the shrimp farming industry.The carbon sector encompassed the Carbon sink variable, which signifies the quantity of carbon stored within the mangrove ecosystem.The model computed alterations in carbon storage by considering fluctuations in mangrove area over time.

Shrimp pond sector
The shrimp pond sector focused on the area of shrimp ponds, which represents the dynamics of shrimp ponds within the system.The model accounted for the rise or fall of shrimp aquaculture area over time and explores the factors influencing this process, focusing on the conversion between shrimp ponds and mangroves.
Besides, the Effect of land availability variable was critical for understanding the dynamics of the pond sector.It measured the impact of land availability on the net pond change over time.It considers the competition for land between mangroves and ponds within the system (Vaz 2014, Van et al 2015, Bosma et al 2016, Osland et al 2018).When land availability is limited due to a large mangrove area, it reduces the space for pond expansion, potentially slowing the net growth of pond areas.On the other hand, if the mangrove area decreases, more land becomes available for pond expansion, promoting the net growth of pond areas (Richards and Friess 2016).

Economics sector
The economic sector explained the relationship between shrimp production and its economic output.Driven by economic incentives, shrimp farmers often aim to expand their shrimp ponds by resorting to deforestation.For example, in Setiu, Terengganu State, Malaysia, economic motives played a important role in the conversion of land to aquaculture (Isa et al 2021).The profit gained from shrimp farming incentivizes farmers to further convert more mangrove areas to shrimp ponds (Arquitt et al 2003, Arquitt et al 2005).Barbier and Cox (2003) discovered that GDP affected mangrove loss in 89 countries from 1985-2000.
The Vietnamese shrimp industry benefits local farmers financially and creates jobs in the shrimp industry.In 2010, Vietnam's aquatic products contributed 4.6% to the GDP, equivalent to US$ 4.8 billion, with shrimp production playing an important role and providing jobs for over 4 million people (Ha 2012).Therefore, the feedback loop in the economic sector explains the relationship between shrimp aquaculture expansion and economic incentives.Higher shrimp production leads to increased Gross Domestic Product (GDP) in the shrimp sector, which creates profitable incentives for shrimp farmers to expand their operations.This, in turn, leads to further pond development, forming a reinforcing feedback loop.
2.1.1.5.Employment sector One of the advantages of shrimp farming development is that it creates many job opportunities (Nila Rekha et al 2015).Especially, more than 60% of the Ca Mau province's rural employed population was engaged in the shrimp industry in 2021 (Ca Mau Statistics Office 2022).Thus, the employee sector in this study focused on measuring the number of people engaging in this industry and the contribution of shrimp farming to rural employment generation.

Loops analysis
From the feedback model, we identified 9 primary loops and categorized them into four 'mangrove' loops and five 'shrimp ponds' loops (table 1).Loop 1 represents a negative feedback loop, indicating that the presence of mangroves influences land availability.As mangrove forests occupy land, they limit the availability of open space for further conversion, contributing to their own preservation.Conversely, loop 2 signifies a positive feedback loop, where mangrove conversion to shrimp ponds initiates a cycle of reciprocal conversion.Initially, mangrove deforestation leads to pond creation, but over time, the ponds that are unprofitable may be converted back into mangrove habitats, promoting ecosystem restoration and resilience.Loop 3 portrays another negative feedback loop, where the effect of land availability on mangroves influences pond expansion.As land becomes available due to mangrove deforestation, it facilitates pond construction, thereby impeding mangrove rehabilitation efforts.Finally, loop 4 represents a negative feedback loop, as mangroves are converted into ponds, the land available for mangrove growth decreases, thereby reducing further conversion of mangroves into ponds.This process aids in maintaining or restoring mangrove areas.These feedback loops underscore the intricate relationship between mangrove conservation and shrimp farming expansion, emphasizing the importance of sustainable land-use practices and ecosystem management strategies to mitigate adverse impacts on coastal ecosystems.
The second group of loops describes the intricate interactions between shrimp farming expansion, land availability, and mangrove ecosystems.Loop 5 illustrates a negative feedback loop, indicating that the availability of land influences the expansion of shrimp ponds.As land becomes scarcer due to pond development, the potential for further pond expansion diminishes, thereby restraining the growth of shrimp farming.In contrast, loop 6 signifies a positive feedback loop, wherein the economic profitability of shrimp farming incentivizes further pond expansion.Higher shrimp GDP generates profitable incentives, encouraging farmers to expand their operations and invest in additional pond construction.Loop 7 represents another negative feedback loop, where pond-to-mangrove conversion influences land availability.The conversion of ponds back to mangrove habitat reduces the available land for shrimp pond expansion, acting as a regulatory mechanism to curb further encroachment into mangrove ecosystems.Conversely, loop 8 depicts a positive feedback loop, as pond-tomangrove conversion makes more mangrove areas available for potential deforestation, leading to an increase in available land for pond construction.This cycle perpetuates the conversion of mangroves to shrimp ponds, exacerbating the loss of critical coastal habitats.Lastly, loop 9 illustrates a negative feedback loop, indicating that the effect of land availability on ponds influences mangrove-to-pond conversion.As land availability decreases, the pressure to convert mangrove areas into shrimp ponds diminishes, mitigating the rate of mangrove loss.These feedback loops underscore the complex interplay between economic incentives, land-use decisions, and  environmental conservation efforts in coastal areas, emphasizing the need for integrated management strategies to promote sustainable development and preserve mangrove ecosystems.

Stock-flow diagram
The stock-flow diagram (SFD) of this paper is presented in figure 2. The SFD includes stock, flow, and auxiliary variables.Firstly, the parameters of the model were identified used historical data from 2007 to 2021; then relevant variables were fitted and regressed to predict future trend until 2050.

Mangrove sector
The mangrove sector comprised the Mangrove stock variable, which represents the total mangrove area.This variable was influenced by flow variables, including mangrove Added area and Lost area, as well as auxiliary variables such as Mangrove planted area, Mangrove lost rate, Mangrove to pond conversion rate, Pond to mangrove conversion rate, and Effect of land availability on mangrove.We calculated the net change in mangrove area over time by subtracting the area of mangroves lost from the area of mangroves added.This equation reflects the balance between mangrove restoration and degradation within the system.If the net change is positive, it indicates a growth in the mangrove area, while a negative net change implies a decline in the mangrove area.

Carbon sector
The carbon sector encompassed the Carbon sink stock, which signifies the quantity of carbon stored within the mangrove ecosystem.The model evaluated the temporal variation in carbon storage by accounting for changes in mangrove area through flow variables denoting Carbon stored and Carbon lost.The Carbon stock per hectare value auxiliary variable (Nam et al 2016) served as a conversion factor for quantifying the carbon content within the mangrove forest.This interaction sheds light on the role of mangroves as carbon sinks and the potential implications for climate change mitigation.

Shrimp pond sector
The shrimp pond sector focused on the stock Ponds, which represents the shrimp pond area.The model measured the rise or fall of shrimp aquaculture area over time through the Net changed pond area flow variable and explores the factors influencing this process.There were auxiliaries that helped explain the dynamic of the shrimp pond area, such as Pond net growth rate, Profit incentive, Mangrove to pond conversion rate, Pond to mangrove conversion rate, and Effect of land availability on ponds.Moreover, the auxiliary variable 'land capacity' imposed a constraint on the expansion of both shrimp ponds and mangrove areas.This growth limitation was set at 464,105 hectares, representing the total area designated for agriculture and aquaculture in Ca Mau province (Ca Mau Statistics Office 2022).Furthermore, the expansion of mangrove area in Ca Mau partially resulted from mangroves, inefficient rice yields, and other land uses (Clark Labs 2022).However, this study specifically concentrated on the interaction between mangrove and shrimp ponds through the pond to mangrove and mangrove to pond conversion variables.Consequently, we presumed that any changes in shrimp pond area not attributable to the conversion between mangrove and ponds were explained by other land uses.

Economics sector
The economic sector revolved around the Wealth stock variable, symbolizing the economic output derived from shrimp ponds.Generally, the shrimp industry's gross domestic product (GDP) is determined by its yield or productivity, number of seasons, and market considerations such as prices, demand, and exporting.However, these variables change depending on the type of shrimp, such as black tiger shrimp and white shrimp, as well as their different sizes.All of these factors require substantial efforts and transcend the scope of the study.Given that the purpose of this article is to examine the shrimp aquaculture area and its associated consequences, we simplified the economic sector and assumed that the production value from shrimp farming can be interpreted via the area variable.A similar approach was utilized to estimate agricultural output value in Sichuan, China (Wang et al 2022).The model calculated the Wealth stock over time based on the Shrimp GDP calculated from the Pond area and Output value per hectare.

Employment sector
The shrimp employment sector encompassed the Shrimp employees stock variable, which describes the number of shrimp farming employees.The model calculated the change in shrimp employees over time based on the Net changed pond area, Area per shrimp farm, and Workers per farm variables.According to the Rural, Agricultural, and Aquatic Census of 2016, the average labor per farm in Ca Mau is 2 people (Department of Agriculture Forestry and Fisheries Statistics 2017).This analysis assumed that the number of laborers per shrimp farm is constant; hence, the number of shrimp employees positively correlates with changes in shrimp ponds over time.
The rural population was explained by the stock Rural population in Ca Mau.The model calculated the change in Rural population over time based on the Rural population growth flow variable and the Rural population growth rate auxiliary variables.Their interaction helps understand rural demographic changes and their potential impacts on the system.
The stock variable Rural employees served as a measurement of the number of employed people in rural areas.The model determined the fluctuations in rural employees over time by considering various auxiliary variables such as Rural population growth, Rural working age rate, and Rural unemployment rate.It enables us to gain insights into how changes in the rural population and employment rates influence rural employed people.
In addition to economic benefits, shrimp aquaculture also creates rural employment opportunities (Ha 2012, Duy et al 2022).Thus, we measured the shrimp industry's contribution in rural employment over time through the ratio of Shrimp employees to Rural employees.
All of the variables and equations are presented in table 2.

Scenarios
The model was first simulated using historical data from 2007-2021 and then predicted until 2050 with a time step of 1.This study examines two scenarios of the interaction between mangrove conservation and shrimp pond expansion in Ca Mau (table 3): (1) The first scenario, referred to as the business-as-usual (BAU) scenario or base scenario, utilized the average change trends of all variables in the models from 2007 to 2021 and projects future trends up to 2050.Additionally, no policy constraints were imposed on the expansion of shrimp aquaculture.
(2) The second scenario, referred to as the Policy scenario, integrated the target policy outlined in the 'Enhancing Efficiency and Sustainable Development of the Shrimp Industry in Ca Mau Province by 2025 and Vision to 2030' project (Ca Mau DARD 2022) into the analysis.In accordance with this policy intervention, the shrimp aquaculture area in Ca Mau was stabilized at 280,000 ha from 2030 onwards by adjusting the related variables, thereby ensuring that the stock of ponds does not surpass this threshold after

Pond to mangrove conversion
Increase by 1.5 times reaching it.Furthermore, we posited that stabilizing the shrimp farming area would result in unused and abandoned ponds being converted back into mangroves, thereby decreasing deforestation from mangroves to shrimp ponds.All other variables followed similar parameters in the BAU scenario.

Model validation
An important task in system dynamics is to attain a certain confidence level to validate a system dynamic model (Duggan 2016).A model gradually accumulates confidence as it passes more tests because there is no single test to validate a system dynamic model.A system dynamic model is validated if it has structural validity and behavioral validity (Barlas 1996).This study conducted 20 different tests to validate the model structure and behavior using the package 'RUnit' (König et al 2007) and instructions from Duggan (2016), wherein Tests 1-18 were for structural validity and Tests 19-20 were for behavioral validity (table 4).In the BAU scenario, the projection indicates a linear expansion of shrimp aquaculture in Ca Mau if it faces no growth constraints.The projected shrimp pond area will reach a maximum of 310,916 (± 4253.313)hectares by 2050.Overall, the BAU scenario reflects a growth-oriented approach with continuous expansion in shrimp aquaculture.Another study in Malaysia with a similar approach predicted a similar increase in shrimp farms over the next 30 years (Isa et al 2021).

Results and discussion
Under the Policy scenario, shrimp aquaculture exhibits fluctuations similar to the BAU scenario from 2021 to 2024.A deviation between the two scenarios will emerge in 2025 as shrimp expansion in the Policy scenario becomes lower than the BAU scenario (figure 3).Furthermore, the net changes in the pond area in the two scenarios also differ.In the Policy scenario, the net changes are smaller in magnitude and tend to diminish over time, implying a more moderate and controlled rate of expansion than the rapid expansion observed in the BAU scenario.Due to policy interventions, the shrimp area is projected to stabilize and be maintained at 280,000 hectares beginning in 2027, in accordance with Ca Mau Province's target policy of sustainable development for the shrimp sector.From 2027 to 2050, the net changes in shrimp farming ponds will be constant at zero, suggesting pond area stabilization aligned with the target policy.Comparing the two scenarios reveals that the Policy scenario generates a distinct trajectory in the pond area compared to the BAU scenario.These divergent trajectories will result in varying implications for the environmental, social, and economic aspects of the following analyses.It is important to note that this study assumed linear growth in shrimp production expansion in Ca Mau due to constraints in the modeling process and the lack of scientific evidence to predict future fluctuations in shrimp pond expansion.Several factors can either promote or constrain the development of the shrimp industry, including government policies, market demands and prices, unforeseen diseases, and environmental conditions.
Changes in the shrimp pond area correlate with variations in the mangrove area.between 2007 and 2021, there was an overall decline in the added mangrove area, with values ranging from 1,666 hectares in 2007 to 998 (± 64.656) hectares in 2021 (figure 4  mangrove areas, such as in 2007, 2008, and 2009.It is notable that the increase in mangrove loss area coincided with the shrimp aquaculture boom.These fluctuations in lost mangrove areas suggest variations in the extent of mangrove degradation or conversion over time due to the expansion of shrimp aquaculture. In the BAU scenario, the continuous expansion of shrimp farming results in further deforestation toward mid-century, as evidenced by the decreasing mangrove new area in figure 4(a).The mangrove lost area from 2022 to 2050 in the BAU scenario averages 1123 hectares per year and shows minimal fluctuation over the projected years (figure 4(b)).Due to the relatively small mangrove area in this scenario, there will be less available mangrove area to be converted into shrimp ponds.Thus, the increases in shrimp ponds in the mid-century might come from other land uses such as agricultural activities.Additionally, given the severe degradation of mangrove areas in Ca Mau since 1995 (Truong and Do 2018), local authorities are expected to implement stringent regulations and monitoring to avoid further deforestation.
In the Policy scenario, stabilizing shrimp farming areas at 280,000 hectares allows mangrove forests to recover.The pressure of land conversion from mangroves to ponds is alleviated, resulting in adding more new areas over the years (figure 4(a)).On average, the projection indicates that approximately 1396 hectares of mangroves will be added annually from 2022 to 2025.Conversely, the lost area of mangroves is lower than the added area, leading to an increase in the total mangrove area of Ca Mau in the Policy scenario towards the midcentury.Besides, the lost area of mangroves in the Policy Scenario gradually surpasses that in the BAU scenario (figure 4(b)), as there would be more mangroves available for timber harvesting.According to land tenure policies, planted mangroves older than 10 years are eligible for timber harvest, providing an important alternative income source for farmers in Ca Mau to offset reduced income from aquaculture (Ha et al 2014).
The variations of added and lost mangrove areas were combined to calculate the mangrove area in Ca Mau  In the BAU scenario, continuous expansion of shrimp ponds in Ca Mau leads to the ongoing loss of mangrove area from 2022 onward until 2050 (figure 5).The projected mangrove area in 2050 is only 70,349 (± 888.801) hectares, which is about 8,962 ha less than the peak level observed in 2014.
Conversely, in the Policy scenario, Ca Mau's mangrove area demonstrates a different trend than the BAU scenario (figure 5).The projection suggests a recovery of mangrove forests in the province over time.The growth rate is faster during the 2022-2030 period and gradually slows down towards 2050.The projected average net change of mangrove area from 2022-2050 is 224 hectares per year.Despite this projected increase, the mangrove area in Ca Mau in 2050 (77,016 (± 687.155) hectares) is still 2,295 hectares lower than the peak observed in 2014.This finding indicates that additional efforts are required, beyond the regulation of shrimp farming expansion, for Ca Mau's mangroves to fully recover.
The disparity in projected shrimp pond and mangrove area in 2050 between the two scenarios is 30,917 hectares and 6,667 hectares, respectively.This means that a 30,917-hectare increase in shrimp ponds corresponds to a 6,667-hectare decrease in mangrove area in 2050.Overall, the Policy scenario shows noticeable and sustainable growth of mangroves over time, while the BAU scenario depicts a gradual decline in mangroves due to the constant demand for land use conversion to expand pond areas.This finding indicates that successfully stabilizing the shrimp farming area indirectly benefits mangrove protection.In Thailand, a 'feebate' policy action could potentially balance effects on mangrove regions (Arquitt et al 2003).However, in the Southern Tamil Nadu coast, India, aquaculture farms present a challenge for coastal management efforts, necessitating careful planning and strategic measures to mitigate the loss of valuable mangrove coasts due to aquafarm expansion (Rajakumari and Rajaram 2021).

Carbon sink
Because this study adopted a constant auxiliary to represent the potential carbon stock per ha (Nam et al 2016), the trend of the stock variable Carbon sink followed the trend of mangrove growth in both the BAU and Policy scenario (figure 6).In the mid-century, the cumulative carbon sinks in the BAU and Policy scenario will reach 59 × 10 6 (± 0.75 × 10 6 ) MgC and 65 × 10 6 (±0.58 × 10 6 ) MgC, respectively.This indicates a difference in the carbon sequestration potential between the two scenarios, implying that controlling and managing shrimp aquaculture expansion will result in a greater carbon sequestration capacity for the mangrove forests of Ca Mau.
It is worth noting that the estimation of carbon stocks in this study serves as a rough indicator for the potential of climate change mitigation via carbon storage of Ca Mau mangrove forests because determining carbon sequestration or carbon storage is extremely complicated and beyond the scope of this study.Mangrove forests are the most important carbon sinks in the tropics (Tue et al 2014).Thus, the quantification of carbon storage in mangrove forests is essential for the development of Vietnam National Strategy For Climate Change

Shrimp GDP
The shrimp industry stands out as a pivotal economic sector, driving the economic growth and rural development of Ca Mau province.The output value of shrimp farming areas averages around 45 million VND/ ha (US$ 1890/ha).Laborers in this sector earned 46.6 million VND (US$ 1680)/person/year (Ca Mau Statistics Office 2010, 2013, 2022).Ca Mau is the largest producer of black tiger shrimp production in the Mekong Delta and operates farms with the highest cost efficiency but low environmental efficiency (Trang et al 2023).
During the baseline period, the shrimp GDP experienced a remarkable increase from 9,227 billion VND (US $ 369.080 million) in 2007 to a peak of 15,848 billion VND (US$ 633.920 million) in 2012 (figure 7).This trend mirrors the shrimp aquaculture boom observed since the 2000s (Son et al 2015).Subsequently, the shrimp GDP declined from 2012 to 2015, followed by a rebound until 2019.In the BAU scenario, following 2021, shrimp GDP continues to increase as the forecasted production area continuously expands, albeit at a slower pace compared to the period from 2007 to 2012.This upward trend persists until it reaches a second peak value of 14,250 (± 0.336) billion VND (US$ 570 million) in 2050.Similarly, aquaculture expansion also led to increased income for farmers in Malaysia (Isa et al 2021).
In the Policy scenario, shrimp GDP follows a different pattern.From 2027 onward, the projected values remain stable at 12,833 (± 0.286) billion VND (US$ 513.320 million) (figure 7), resulting from capping shrimp ponds at 280,000 ha to meet the policy target.This indicates a drawback of the managerial target of shrimp aquaculture in Ca Mau.The cap on shrimp aquaculture may hinder the industry's economic growth and limit the sector's ability to generate additional revenue in the long term.While the expansion constraint aims to address environmental concerns, it is crucial to find a balance between sustainability and economic growth in the shrimp industry.Moreover, when designing policies, it is important to consider the production scale and the livelihoods of smallholders.Shrimp farming, being capital-intensive, allows large entrepreneurs to easily engage in high-profit activities, while small-scale producers would require institutional, technical, and financial Various strategies should be considered to promote the shrimp industry's economic growth in the Policy scenario.These include improving productivity, enhancing production quality, adopting advanced and hightech farming models, and strengthening production linkages.By carefully considering and implementing appropriate measures, Ca Mau may secure the long-term success and development of its shrimp industry.One of the emerging solutions in Vietnam and Indonesia is the mangrove silvo-aquaculture practice (particularly integrated mangrove-shrimp farming).The implementation of mangrove-shrimp systems has the potential to enhance resilience in coastal aquaculture landscapes by offering opportunities for livelihoods, restoring mangroves, sequestering blue carbon, reducing disease risks, improving water quality and generating additional income (Hochard et al 2019, Akber et al 2020).

Shrimp employees and the contribution in rural employment generation
During the baseline period, the variation of shrimp employees mirrored the boom and bust phases of shrimp aquaculture in Ca Mau (figure 8(a)).In 2007, 379 thousand people were engaged in shrimp farming in Ca Mau.This number increased to 402 thousand people in 2015 before falling to 390 thousand people in 2021.Alongside the growth of shrimp aquaculture, its role in rural employment generation increased, with this sector providing work for 63%-70% of the rural employees from 2007 to 2021 (figure 8(b)).
In the BAU scenario, the expansion of shrimp ponds until 2050 creates more job opportunities and enhances the industry's role in rural employment generation.By 2050, it is projected that the shrimp farming sector will account for 78.9% (± 0.013) of the employed rural population in Ca Mau (figure 8(b)).However, it is important to recognize that the heavy reliance on shrimp farming for employment could result in socioeconomic challenges.Shrimp farming is associated with high risks, and farmers may experience economic losses due to diseases or water pollution (Pham 2008).
On the other hand, the number of shrimp workers in the Policy scenario was projected to be capped at 402 thousand people due to the stabilization of the shrimp pond area (figure 8(a)).Nevertheless, it still contributes a high share to the rural labor force in Ca Mau.In 2050, the projection shows that 72% (± 0.007) of the employed rural population will be working in the shrimp industry (figure 8(b)).
The contribution in rural employment generation in the Policy scenario is lower than in the BAU scenario, and the disparity grows over time.In 2050, the difference in shrimp employees between the two scenarios is 39 thousand people, which is equal to 6.9% of the employed rural population.With fewer work opportunities in shrimp farming, labor migration from rural areas to cities is more likely to increase as individuals seek alternative employment options.These findings reveal further policy challenges and implications for Ca Mau's rural employment.Policymakers should consider strategies to address the contribution of the shrimp industry to rural employment and explore alternative avenues for sustainable job creation in rural areas.Conservation policies, typically aimed at preserving areas of high biodiversity, should also consider the livelihood needs of local communities (Idrus et al 2019).Implementing these policies may need to be gradual while simultaneously promoting alternative livelihood opportunities (Akber et al 2020).

Conclusions
This study sheds light on the interactions between mangroves and shrimp aquaculture in Ca Mau and forecasts future development under different scenarios.Since the 2000s, the rapid expansion of shrimp aquaculture has put pressure on land use management in Ca Mau.Consequently, shrimp ponds have replaced a remarkable portion of Ca Mau's mangrove forest, preventing the mangrove area from recovering to its peak in 2014.The expansion of shrimp farming also compromised the potential for carbon storage in the mangrove forest.Despite its unfavorable environmental impacts, shrimp aquaculture is critical to the province's economy and rural employment generation.
According to projections derived from the system dynamics model, under the BAU scenario, if no intervention occurs and shrimp aquaculture in Ca Mau follows its current trajectory, the shrimp farming area will continuously expand until 2050.This expansion will further drive the conversion of more mangrove areas into ponds, resulting in a reduction of the mangrove area in Ca Mau to 70,349 (± 888.801) hectares by 2050.Consequently, the potential for carbon storage in mangrove forests will decrease.However, the ongoing development of shrimp aquaculture will yield economic advantages, contributing 14,250 (± 0.336) billion VND (US$ 570 million) to the province's GDP by 2050.Furthermore, the expansion of shrimp farming will create additional employment opportunities for 441 thousand people and help sustain rural livelihoods.
In contrast, under the Policy scenario, implementing the policy objective of capping shrimp aquaculture at 280,000 hectares by 2027 will yield different implications for Ca Mau's environment, society, and economy.Implementing a cap on shrimp pond areas will mitigate the future conversion of mangroves into shrimp ponds, facilitating gradual regrowth of the mangrove forest over time.By 2050, the Policy scenario forecasts an additional 6,667 hectares of mangrove forest compared to the BAU scenario.This scenario also predicts an improvement in carbon storage, indicating the potential of managed shrimp farming areas to contribute to climate change mitigation efforts.However, concerns arise regarding the social and economic consequences that authorities must address.First, assuming all other economic factors remain constant, constraining shrimp aquaculture could lead to stagnant economic growth and potential conflicts with Ca Mau's other development objectives.Second, the shrimp farming sector may generate fewer employment opportunities in rural areas, resulting in indirect consequences such as unemployment, migration, and insecure rural livelihoods.It is imperative for the government to thoroughly analyze these potential effects and implement necessary adjustments accordingly.
In reality, achieving sustainable growth in managing mangrove forests and shrimp aquaculture in Ca Mau necessitates local authorities to consider multiple factors with a broader scope beyond simply controlling the shrimp pond area at 280,000 hectares in the near future.This study proposes several policy considerations.First, given the pressures on land use resulting from the rapid expansion of shrimp aquaculture, it is crucial to develop and implement effective land use management policies in Ca Mau.This entails striking a balance between the needs of shrimp farming and the preservation and restoration of mangrove forests to ensure their long-term sustainability.Implementing measures to prevent further conversion of mangrove areas into shrimp ponds and promoting mangrove rehabilitation initiatives are essential for maintaining biodiversity, enhancing carbon storage, and mitigating the negative environmental impacts of shrimp farming.Second, although shrimp aquaculture substantially contributes to the province's economy, diversifying economic activities in Ca Mau is still necessary, especially in rural areas.Encouraging the development of other industries and sectors can reduce dependence on shrimp farming and mitigate potential risks associated with stagnated economic growth if shrimp aquaculture is capped.Third, to address the potential loss of employment opportunities in rural areas resulting from capped shrimp aquaculture, the government should promote alternative livelihood options.This can include supporting the development of other agricultural sectors, promoting rural entrepreneurship, and encouraging participation in payment for carbon services.
The methodology outlined in this study extends beyond Ca Mau and is applicable to other regions facing threats to mangrove ecosystems due to aquaculture expansion.Such areas include the Mekong River Delta, the Red River Delta, as well as specific provinces like Hai Phong, Thai Binh, Nam Dinh, Thanh Hoa, Quang Nam, Quang Ngai, Phu Yen, Ho Chi Minh City, Bac Lieu, and Kien Giang.While the methodology is transferable, the outcomes may vary depending on local policy objectives or applied scenarios.Moreover, the SDM could be scaled up to address broader coastal management challenges, particularly in Southeast Asia.
While the study made efforts to analyze the intricate system of mangroves and shrimp aquaculture, it encountered limitations due to the complexity of certain variables and a lack of extensive historical data.This study assumed that the trends of certain variables followed linear trends and did not account for variables such as inflation over time, market dynamics, unforeseen economic events, or future climate extreme events, all of which could have influenced the projections of shrimp farming areas and the growth of mangrove forests.

Figure 1 .
Figure 1.Feedback diagram for shrimp aquaculture expansion and mangrove conservation.

Figure 2 .
Figure 2. Stock-flow diagram for shrimp aquaculture expansion and mangrove conservation.
(a)).This indicates a gradual decrease in the rate at which new mangrove areas are being added over time as mangroves are converted into shrimp ponds.In contrast, the lost mangrove area in Ca Mau exhibited a fluctuating pattern with no clear overall trend (figure 4(b)).The values fluctuated from 329 hectares in 2007 to 3,560 (± 192.557) hectares in 2021.There were years of relatively low loss of
from 2007-2050 (figure 5).During the period 2007-2014, the total mangrove area of Ca Mau continuously increased from 70,072 ha to 79,310 ha (Clark Labs 2022) before declining to 75,424 ha in 2021.The consistent growth from 2007 to 2014 came from the efforts to meet the target of the National Biodiversity Strategy to 2010 and Vision to 2020 (Government of the Socialist Republic of Vietnam 2007).Additionally, diseases and water pollution negatively affected shrimp aquaculture during this period (Anh et al 2010, Bosma et al 2016).Most critically, the enormous expansion in shrimp aquaculture beginning in 2014 coincides with the start of the intensive destruction of Ca Mau's mangrove forests.This demonstrates a direct relationship between these two land uses.Because of the tremendous economic prospects of the shrimp industry, a considerable portion of the mangrove forests in Ca Mau were turned into shrimp ponds.Ca Mau mangrove forest experienced remarkable losses after reaching its peak in 2014 and has yet to recover, despite subsequent conservation efforts.Previous

Figure 8 .
Figure 8. Shrimp employees and their contribution in rural employment generation 2007-2050.

Table 1 .
Feedback loops and their signs.
Data for this research was sourced from Ca Mau Statistical Yearbooks spanning 2007 to 2021 and Agricultural and Fishery Censuses from 2006 to 2016.Clark Labs (2022) provided spatial data to evaluate mangrove changes in Ca Mau.

Table 2 .
Variables and equations in the model.