Multi-Scale Drivers of Land-Use Changes at Farm Level II: Application of Conceptual Framework in the Salinity Intrusion Zone of the Vietnamese Mekong Delta and Cross-Case Comparison with the Highly Flooded Zone

Abstract

Frequent drought and worsening salinity intrusion challenge future land uses and livelihoods in the Vietnamese Mekong Delta. The central government is, therefore, formulating a new strategy premised on adaption to natural environmental dynamics. For an achievable strategy that bridges the gap between plans and practice, it is important to understand what drives land-use changes at the farm level. Previous research developed and applied a multi-scale drivers framework in the highly flooded zone of the delta. The current study uses that same framework to investigate the land-use history and drivers of change in the salinity intrusion zone of this delta from 1975 to 2016. We interviewed 32 farmers in Tra Vinh Province and used transcript analysis to quantify the influence of the drivers that the farmers mentioned. We then compared the drivers of land-use change with those found earlier in the highly flooded zone. Results show more diversification of land uses and land-use changes in the salinity intrusion zone. Farmers here followed three main pathways: rice intensification, integrated farming of rice and vegetables/aquaculture, or intensive shrimp farming. Land-use changes were conditional on the regional infrastructure construction to preserve freshwater conditions. However, household-scale drivers, especially natural and financial assets, were most frequently mentioned. Socio-economic context also emerged as an important driver, particularly trends and pressures from the community and markets.

1. Introduction

The Vietnamese Mekong Delta (VMD) plays an essential role in Vietnam’s food security. Farmers in the delta produce some 55% of the nation’s rice output and nearly 70% of its aquaculture products [1]. Nonetheless, the delta is particularly vulnerable to the impacts of climate change and sea-level rise [2,3]. These, combined with human interventions, particularly dam construction upstream, have sharply reduced freshwater flows into the delta [4]. Shortages of freshwater have allowed saline water to intrude further inland, endangering agriculture in the year-round freshwater zone [5,6]. In the coastal estuary of the VMD, increasingly severe salinity intrusion and more frequent droughts are now a grave concern, and great losses in agricultural production and livelihoods have been recorded in recent years [6]. Moreover, natural flows and hydrological regimes of the delta are changing in unexpected ways due not only to the expansion of irrigation works [7] but also to sand mining [8,9] and groundwater extraction. The result is subsidence and riverbank erosion, alongside worsening salinity intrusion [10,11,12].

To cope with the combined effects of climate change and human influences, new visions for land use and the infrastructure system of the VMD are urgently needed. The central government has taken on this challenge, reforming national food security policy and setting new objectives for delta development [13,14]. A main goal is to restructure agricultural production, embracing higher quality and higher value products such as sustainable brackish and saline aquaculture and fruits, as well as rice production. Government strategy, thus, foresees a transformation of land and water use in the delta toward greater agricultural diversification and “living with salinity intrusion” [15] (p. 3). In this regard, local authorities and farmers in the salinity intrusion zone still struggle to find compromises to solve land-use conflicts and allow for sustainable future-oriented farming systems [16,17,18]. For feasible plans tailored to the different delta zones, policymakers need better, more localized information on the multi-scale drivers of land-use change as well as the role of water infrastructure interventions resulting in the trends of land-use changes in the delta.

Numerous studies have examined the impacts of climate change and human interventions on the VMD. These investigate the relationships between the multiple drivers of land-use change based on different conceptual frameworks and data from literature reviews, key informant interviews, focus groups, and household surveys. Some offer recommendations and solutions to help farmers cope with extreme events. For example, Tran et al. (2018) [19] used multiple logistic regression models to investigate factors affecting inhabitants’ knowledge, attitudes, and practices regarding land use in Vietnam’s southernmost province of Ca Mau. They found that land-use changes at the household level were influenced mainly by “educational level, age, participation in a group activity, access to communication means, and occupation”. Their conclusions emphasize the need for sufficient comprehension of local drivers of land-use change before implementing land-use management projects. Nguyen Thanh et al. (2021) [20] found that coastal farmers nowadays prefer to diversify their crops and livelihoods. They summarized the key drivers of land-use change in three provinces (Ben Tre, Soc Trang, and Kien Giang) as the ‘4Ps’ (price, producer, place, and policy) and proposed a sustainable agricultural transformation framework for the VMD coastal estuary zone. In a study in Soc Trang and Kien Giang provinces, Nguyen et al. (2019) [21] adopted complex adaptive systems theory to investigate the drivers of land-use change. Their results indicate that the agricultural transformations of past decades were determined by “policy intervention, farmers’ desire for profit maximization, changing salinity conditions, and technological development”. They observed, however, that as salinity intrusion becomes more severe, coastal farmers might choose problematic adaptations that ‘lock’ agricultural production into particular systems or constrain change in desired directions. The study concludes that adaptation pathways should be chosen to allow for continuous and reversible change.

A common thread running through these studies is the recognition that land-use change at the farm and community level is driven by multiple factors that are reciprocally related and operate at different scales [22,23,24,25]. Therefore, to understand the drivers of land-use change in the VMD, we developed a multi-scale drivers framework consisting of 13 key factors active at three scales [26]. The framework builds on two existing works, namely, the micro-level drivers of the land-use change framework [24] and the on-farm decision-making process [25]. To these, we added drivers derived from an analysis of interview data using the grounded theory method [27]. The framework is elaborated in more detail in a companion article [26], which also presents findings from its initial application to investigate land-use changes in the highly flooded upper delta zone.

The current article presents a second application of the framework in a very different agro-hydrological setting: the salinity intrusion estuarine zone. Here, opportunities and limitations for land-use change are determined largely by soil characteristics, freshwater shortages, and salinity intrusion dynamics. Moreover, government policies and infrastructure development here have long favored freshwater rice cultivation, though these efforts have produced mixed results and, in some cases, proven to be incompatible with climate change trends [3,6]. The VMD estuarine zone is unique in offering farmers opportunities to switch from and between permanent freshwater cultivation, intermittent freshwater/brackish cultivation, and permanent saline water/brackish cultivation [20,21,28,29]. Such switches, however, have implications for the soil and water management regime.

To bridge the divide between the government’s land-use and water management plans and their practical implementation, a comprehensive analysis of land-use changes and their driving factors at the farm level is a necessary contribution. Hence, this research applied the multi-scale drivers framework to tackle the subsequent inquiries: (1) What patterns characterize the pathways of land-use change within the salinity intrusion zone of the VMD? (2) What primary factors significantly steer land-use decisions at the level of individual farms? (3) To what extent do different socio-hydrological conditions within the delta region shape the drivers of land-use changes and subsequently mold the resulting land-use change patterns? The third research question will be addressed through a comparative analysis of the outcomes of our earlier investigation conducted within the highly flooded zone of the VMD [26]. By contrasting results from distinct agro-hydrological case studies, we can derive more comprehensive conclusions regarding the intricate interplay among land-use change, water resource management, and various driving factors inherent to delta management.

Furthermore, our application of the multi-scale drivers framework in the salinity intrusion zone serves a dual purpose. Firstly, it serves to assess the validity and rigor of the framework itself. Secondly, it contributes to a deeper comprehension of land-use changes and their driving factors across various agro-hydrological zones in the VMD. The examination of the driving factors behind land-use changes, as perceived by local farmers in our two case studies, provides a valuable medium for evaluating and discussing the significance of factors operating at smaller scales. This perspective contrasts with the broader national policy and its consequent water interventions.

Section 2 of this article introduces the salinity intrusion zone case study and methods of data collection and analysis. Section 3 presents the case study results and a cross-case analysis [30] comparing the results from the highly flooded upper delta zone [26] to those from the salinity intrusion zone. Section 4 presents discussions on the application of the multi-scale drivers framework [26]. Section 5 presents conclusions.

2. Case Study in the Salinity Intrusion Coastal Estuary

2.1. Study Area Characteristics

We conducted the study in Tra Vinh Province due to its agro-hydrological characteristics [31], which mirror the alternating cycles of fresh and saline water conditions observed in the salinity intrusion zone of the VMD [20,21,28,29]. Within this context, local farmers have the flexibility to transition between various farming systems. Moreover, Tra Vinh Province benefits from the protection provided by the South Mang Thit salinity prevention infrastructure system [32,33]. This significant water infrastructure initiative, overseen by the central government, plays a pivotal role in shielding coastal regions of the VMD that engage in freshwater agriculture. By delving into the historical trajectory of land-use changes in Tra Vinh Province, we can discern disparities in the driving factors behind these changes prior to, during, and subsequent to the establishment of the salinity prevention infrastructure. This investigation enables us to comprehensively examine how this critical intervention has shaped the dynamics of land-use change in the salinity intrusion zone.

Salinity intrusion in Tra Vinh province is determined by the waters of the Mekong River system and the tidal regime of the East Sea. The main constraints to farming here are salinity intrusion and the acid sulfate soils, especially in the dry season. Farmers make the most of the limited rainwater for rice cropping in the rainy season. Natural conditions are favorable for a rice–shrimp farming system; that is, rainfed rice is cultivated in the monsoon season, followed by brackish shrimp cultivation in the dry season. As in much of the VMD, government policy and infrastructure development here have sought to extend and expand rice production. Particularly, an intricate system of canals, dikes, and sluicegates has been constructed to allow double and triple rice cropping. The province’s three main sources of freshwater are the Bassac River to the west, the Co Chien River to the east (a tributary of the Mekong River), and the Mang Thit River to the north (Figure 1). To the south, the province is bordered by the East Sea, from which saline water easily intrudes inland through the river mouths. Lower dry season discharge from the Bassac and Co Chien rivers enables saltwater to penetrate north, also affecting the western and eastern parts of the province [31,32,33].

Due to the increasing difficulty of guaranteeing sufficient freshwater, some farmers are actively shifting to saline land uses in the dry season. In addition, many farms are situated on dune ridges, a coastal terrain where rainwater may be found stored in small, fresh groundwater lenses that float on top of the saltwater aquifer [34]. These freshwater pockets make these traditionally ‘poor’ lands attractive in the increasingly saline dry-season environment. Farms on the dune ridges can access the freshwater lenses for irrigation, while fields on the surrounding alluvial plains are saline.

Another major factor affecting agriculture in Tra Vinh Province is the South Mang Thit project, consisting of a closed system of canals, dikes, and sluicegates built from 1993 to 2004. The project extends across most of the province, with the aim of enhancing freshwater supply and protecting the upper and middle parts of the province from salinity intrusion. The province today has more than 600 km of large canals and 2000 km of smaller canals. Waterways extending southwards play a significant role in distributing freshwater from the Mang Thit River both inland and toward the coastal areas (Figure 1) [31,32,33].

Salinity intrusion and freshwater scarcity nonetheless pose an increasing threat to the province. Severe drought occurred each year from 2016 to 2020, and shortages of irrigation water created tough conditions for freshwater agriculture. Salinity intrusion is encroaching further inland, even penetrating beyond the sluicegates, not only through those that are opened for shrimp farming in the south but also those meant to protect freshwater rice farming in the north [16]. Shifting rainfall patterns create a salinity intrusion threat in the dune-ridge agro-hydrological system, too, as storage of rainfall in lenses is less efficient when dry spells lengthen. To adapt to smaller freshwater lens sizes, farmers have shifted to vegetables with a shorter growing period.

2.2. Data Collection and Analysis

To enable a cross-case comparison, we followed the same method for our salinity intrusion zone case study as that used in the highly flooded upper delta zone [26]. Applying the multi-scale drivers of land-use change framework, we identified the principal drivers of land-use changes at the farm level in the coastal estuary zone over the 40 years from 1975 to 2016. The framework considers four external drivers of change at the global or international scale (population change, global markets, policies, and climate change) [24,26,35] and 13 drivers of change operating at the national, regional, and household scales (policies, infrastructures, technologies, institutions, markets, socio-economic context, environmental context, social assets, financial assets, human assets, natural assets, physical assets, and willingness) [24,25,26,36] (See Appendix A,

Table A1. List of driving factors mentioned by farmers in Tra Vinh province, Vietnamese Mekong Delta.

Driving Factors	Sub-Driving Factors	Explanations Based on the Interviews
(A) Policies	A1. Collectivization policies (1975–1986)	—Farmers worked in cooperative groups that were organized by the government.
		—The free market was not allowed. The government bought and allocated all products.
	A2. Food-security policies (from 1986 to present)	—The policies aimed to increase the annual rice production to feed the people, export, and earn more profit from rice farming (focusing on quantitative production). Farmers, therefore, were encouraged to intensify their cropping seasons.
		—The state government controls the price of rice production and supports the government’s rice companies through their policies.
		—The state government invested in new infrastructures and agricultural services to support intensive rice farming.
	A3. Natural disaster prevention and control policies	—The state government built and enhanced infrastructures, such as concrete dikes and elevated roads, to protect the lives and properties of people from flood hazards.
		—The army participated in public works to help people when disasters occurred (severe floods, broken dikes, drought, etc.).
(B) Infrastructures	B1. Canals (water supply and drainage system)	—The government operated public works to dredge and enlarge the existing canal networks, aiming to (1) release floodwater faster, (2) provide more freshwater for irrigation, and (3) reclaim lands from acid sulfate soils.
		—The government dredged new canals in the remote areas.
	B2. Low dikes	—The local government and farmers worked together to build and maintain low dikes to delay flooding times or sea tides.
	B3. Concrete dikes also function as concrete roads	—The government built new high dikes or elevated the existing low dike systems to completely protect areas from flood hazards.
		—The government built new national and regional roads, which also function as high dikes.
	B4. Sluice gates and the South Mang Thit system	—The government built a series of sluicegates to control salinity intrusion in the dry season.
	B5. Others (bridges, electricity, domestic/fresh water supply, etc.)	—The government improved the small roads and built concrete bridges.
		—The government improved the freshwater supply system for people in remote areas.
		—Farmers extracted groundwater or stored rainwater for domestic use.
		—Farmers bought solar electricity equipment to generate electricity for pumping and daily use.
(C) Technologies	C1. Farmers improved the physical conditions of their farms	—Farmers used fresh water to treat acid sulfate soils. —Farmers used pumped freshwater from canals or stored freshwater in their ponds to irrigate crops. —Farmers used pumped saltwater in their shrimp ponds. —Farmers used fertilizer, agrochemicals, and organic materials to improve soil quality. —Farmers changed to new rice/vegetable varieties that are more adaptable to their farming conditions. —Farmers applied new techniques to improve the physical conditions of their farms.
	C2. Financial and technical support from the government	—Farmers tried new rice varieties introduced by the government. —Farmers participated in technical training and tried high-value farming models introduced by the local government. —If there were a funded program to try new farming models, the farmers who participated would receive financial and technical support.
	C3. Farmers participate in high-value farming models	—Farmers participated in technical training and tried high-value farming models introduced by the private sector or their friends/family.
(D) Institutions	D1. Cooperative groups	—After the war, farmers had to hand over their lands to cooperative groups operated by the government. They worked together in cooperative groups to produce rice and participate in public works (dredging canals, building dikes, pumping, etc.).
		—Nowadays, farmers do not have to work in cooperative groups, but this kind of group still exists and is in charge of pumping and some agricultural services.
	D2. Local officers gathered the consensus of farmers	—Local officers visited farmers or invited them to community meetings to introduce policies.
		—Local officers organized frequent meetings, trying to persuade farmers to follow the state government’s policies as much as possible.
		—Local officers gathered the votes of the majority. When the government obtained the majority of farmers, they could conduct a project in the commune.
	D3. Local officers operated public works and developed infrastructures	—The local government received funding from the state government to invest in building new high dikes, canals, and roads.
		—The local government is in charge of maintaining the canals and dikes, but the farmers must pay fees for those public works.
	D4. Financial and technical support program launched by the government	—The local government introduced high-yield rice varieties to farmers. —The local government looked for high-value farming models that met the government’s plans and offered farmers technical training. —Agricultural officers introduced farmers to the rice companies or funded projects.
	D5. Introduce farmers to high-value farming models	—Local authorities cooperate with companies and organizations to introduce farmers to high-value farming models.
(E) Markets	E1. Agricultural services	—Farmers paid fees to build/maintain dikes and canals or to pump water in the rainy season.
		—Conditions to rent agricultural machines and labor.
		—Conditions for transportation and storage.
	E2. Traders and markets	—Trends and demands of agricultural products in the market.
		—The confidence between farmers and traders. For example, traders can ask farmers to grow a certain type of rice/vegetables, or they can sign a contract with a secured/fixed price before the start of a cropping season. However, sometimes traders do not follow the contract, they do not come back to buy the products, or they only buy if farmers reduce the price because the price in the market has dropped dramatically.
		—Conditions to do business with traders. For instance, traders prefer to buy products from the major farmers who produce the same type of crops. Those farmers have more advantages in transporting, storing, and selling than small farmers.
		—Conditions to do business with agricultural agents or agricultural extensions. For example, farmers can pay for input products (e.g., fertilizers, agrochemicals, varieties, etc.) after finishing the harvest.
	E3. Price of agricultural products	—The fluctuation of prices in the market.
		—The differences in prices among agriculture/aquaculture products (e.g., vegetables vs. rice, chili vs. watermelon, fish vs. rice/vegetables, etc.).
		—The differences in prices when doing business with rice companies (secured/fixed prices) compared with traders/open markets (fluctuating prices).
	E4. Household expenses	—Everything nowadays must be bought from the market, even fish or vegetables; living expenses, therefore, have increased. That makes farmers want to earn more profit from their farming.
		—Farmers want their children to study in college, hoping the next generation can find a non-farming job. To pay the school fees, farmers need to earn more profit from farming.
		—Middle-aged farmers are not interested in secondary jobs. By increasing the number of rice crops, they expect to earn more income without expending too much labor.
		—If the economic condition of the household is stable, farmers feel satisfied with their farming system. They do not want to invest in a new farming system.
(F) Social-Economic context	F1. Opinion of the majority group/pressure from the community	—For decisions related to policy, an individual farm has to follow the major group in the community (the opinions of 60% of farmers).
		—The farmers who own large lands or succeed in farming have an influence on others’ opinions.
	F2. Neighborhood effects/success or failure of someone in the community	—When farmers see the success of others (neighbors, relatives, other communes, etc.), they are motivated to change their farming system.
		—When many surrounding farms change to a new land use, a farm has to follow the neighbors because those farms influence each other in terms of water usage, pest control, or selling products.
	F3. Immigration	—Farmers move out or move back to the commune due to the changes in land ownership policies. They farm in the commune but live in another commune/their original hometown.
		—Farmers move to the commune because they inherit land from their parents or when they get married to a local (the wife’s hometown).
		—Farmers move to coastal communes to work in shrimp farms.
		—Poor farmers move to remote areas where they can afford more land.
		—Farmers from other regions bring their experiences and ideas about new types of land use to the commune.
(G) Environmental context	G1. Local weather and flooding regime	—Changes in local weather (rainfall pattern, sunny days, temperature, etc.) affect all farms.
		—Salinity intrusion affects the communes on the edge of the south, east, and west.
	G2. Agriculture zone (bounded by canals and dikes)	—Farms that are bounded by the same dikes, canals, roads, and sluice gates share the same infrastructure, water, and flooding conditions.
		—Farms in the dune ridges have specific conditions.
		—If the farms are in an agricultural zone, farmers have to practice a certain farming system that was planned by the government (mostly at the provincial level). For example, large-scale vegetable or fish farms are not allowed in the triple rice zone.
	G3. Farming schedule	—Agricultural activities (transplant, pump water, harvest, spray agrochemicals, etc.) need to follow the same schedule. For example, a farmer cannot irrigate his rice field if his neighbors are going to harvest rice or vegetables.
(H) Natural Assets	H1. Landform (high or low, flat or rough)	—Flat lands with low elevation are easily inundated when it rains and are not suitable for growing vegetables but can be used to grow rice.
		—Lands with high elevation or in dune ridges are feasible for growing vegetables.
		—Lands with rough surfaces require many work hours to be useful for agriculture.
	H2. Soil quality	—The concentration of sulfate acid in soils can be reduced or increased depending on the available freshwater. Sulfate acid is more active in the dry season than in the rainy season.
		—The quality of alluvial soils degrades due to less flooding time (less sediment), salinity intrusion, intensive use of agrochemicals, over-exploitation, or being converted to shrimp ponds.
		—The quality of alluvial soil can be improved with fertilizers or crop rotation.
	H3. Water availability	—Without canals, farmers do not have enough freshwater to treat the sulfate acid soil or to irrigate their crops. Their farming depends on rainwater. Thus, they only grow rice in the rainy season. In the dry season, they grow vegetables.
		—Farmers open the sluice gate or pump freshwater from the canals to treat the sulfate acid soil and to irrigate their rice/vegetables.
		—Farmers open the sluice gate or pump saline water from the canals to operate their shrimp farms.
		—Too much water due to heavy rainfall will cause damage to production. Thus, farmers have to pump water out of their fields.
(I) Physical Assets	I1. Location (close or far from roads and canals)	—Farms that are close to high dikes or roads have advantages in transportation, water use, etc., but they had to give up part of their land when those infrastructures were built.
		—Farms that are close to main and sub-canals use more water resources.
		—In farms close to high dikes/roads, farmers can use the side of the dikes/roads to grow vegetables because this part has a high elevation, and it is easy to transport vegetables.
		—Lands in dune ridges are feasible for growing vegetables.
		—If farmers have many properties that are not located in the same commune, farmers prefer to work on the land near their house; the remote one can be rented.
	I2. Land size (large or small)	—Farmers who have a small amount of land (less than 0.5ha) prefer a high-value farming system such as vegetables, mixed farming, or renting more land to grow rice.
		—Farmers who have large lands prefer the rice farming system because it requires less labor, the agricultural services are convenient, and the market fluctuation is less risky.
		—Farmers who have many lands can try different land-use types or keep their lands for rent.
	I3. Physical conditions were improved by farmers	—Sulfate acid soils were treated. —Pumping freshwater in canals or groundwater to irrigate the crops. —Storing freshwater in canals/ponds to use in the dry season. —Fertilizers, agrochemicals, or organic materials were used to improve soil quality. —Pumping saline water into their shrimp ponds.
(J) Social Assets	J1. Social network	—Farmers who also work for the local government or have relationships with local officers can have advantages in updating information/policies or have a bigger role in the community.
		—Farmers can have the confidence/ trust of the community.
		—Farmers can travel to many places to see and learn from other communes and districts.
		—Farmers have close relationships with traders, companies, and agricultural extensions; some farmers are also traders, service providers, or agriculture agents.
		—Farmers can work together with their siblings, relatives, or cousins, or they can learn new things from family members living far away.
(K) Financial Assets	K1. Extra incomes/other livelihoods	—Farmers can have other livelihoods that contribute to the household income; when they travel to other regions to do other jobs, they can observe and learn new farming systems.
		—Children have jobs, send part of their salary to their parents, and contribute to the household’s income.
		—By increasing the number of cropping seasons, farmers can have extra income.
	K2. Property or household capital	—Economic status of the household (wealthy or poor, do they have savings in the bank).
		—If the agricultural equipment belongs to the household, the farmer can use it or make money by leasing it. Alternatively, farmers have to rent equipment and share them with other farmers.
		—Lands are property; farmers can lease or mortgage their fields to earn income/or capital.
		—The ability to access loans from the bank and financial support from the government.
		—The ability to borrow money from their network (friends, relatives, agricultural agents, mortgagees, etc.).
		—To be in debt is a pressure on farmers. They have to extend and intensify their farming, for example, by triple rice cropping. It is risky for them to change to vegetables because this requires more investment, while the market for vegetables fluctuates significantly.
	K3. Production or profit earned from the farming system	—The difference (lower/higher) in production between two farming systems, leading to a difference in profit. Farmers consider this difference before changing their farming system. They are usually motivated when they see a chance for higher production and profit.
		—Farmers prefer a farming system that has a stable production and profit (the production and the profit depend on both environmental conditions and the market).
(L) Human Assets	L1. Age	—Age and health conditions affect the practice of the current farming system.
		—Age affects the willingness to try new farming systems.
		—Age affects the capacity to learn new farming techniques.
	L2. Education/knowledge/experience	—Farmers with reading ability can learn farming techniques from books and magazines.
		—Farmers can gain farming knowledge from magazines, television, workshops, etc.
		—Farmers with much experience in farming are confident in their current farming system.
	L3. Farming history/system memory	—Farmers keep the same farming system because it is a tradition of the family.
		—Special events in farmers’ lives can cause them to change to a new farming system (e.g., get married and move to the wife’s family).
		—Farmers who have tried many land-use types or livelihoods are more active and confident in changing their farming system.
	L4. Family structure/family labor	—The number of children in the family.
		—Age, education level, and marriage status of the children.
		—The number of people in the family using that farming system.
		—The main laborer who can decide which farming system to use for the household.
(M) Willingness	M1. Interests	—Farmers can prefer a certain farming system and want to keep using that farming system.
		—Farmers have self-motivation to try new farming systems.
		—Farmers can prefer to follow the major farming system in the community.

).

To trace the history of land uses and identify influential factors at particular points of change, we conducted 32 semi-structured interviews with farmers in six communes that exhibit the variety of hydrological characteristics and land uses found in Tra Vinh Province (see Figure 1 and

Table 1. Land uses and demographic characteristics of the interviewed farmers.

Districts	Cang Long	Cau Ngang			Duyen Hai
Communes n: number of interviewed farmers	Binh Phu (n = 5)	My Hoa (n = 4)	Vinh Kim (n = 5)	My Long Nam (n = 7)	Ngu Lac (n = 5)	Don Xuan (n = 6)
Age (years) m = mean (minimums and maximums)	m = 53 (40–64)	m = 54 (48–64)	m = 54 (42–62)	m = 50 (40–60)	m = 52 (34–67)	m = 47 (36–56)
Female household heads	0	0	0	1	0	1
Farm size (ha) m = mean (minimums and maximums)	m = 1.08 (0.6–1.9)	m =1.1 (0.5–2)	m =1.16 (0.6–1.8)	m =3.71 (0.26–17)	m =1.51 (0.5–2.9)	m =1.06 (0.25–2)
Number of farms inside the sluicegate system	5	4	4	7	5	5
Farming systems (number of farms)
Fruit cultivation	1
Triple rice cropping	2					1
Triple rice + vegetables	2				2 (+1) *	3
Double rice cropping				1
Double rice + vegetables			2	2	1
Rice + shrimp farming		1	2
Extensive shrimp farming					2	1
Intensive shrimp farming		3	1	4 (+2) *		1

* Some interviewees had farms at different locations so they could practice two land-use types at the same time.

). Prior to embarking on the interviews, a comprehensive review of pertinent literature concerning geographical conditions and land-use change within the salinity intrusion zone [3,6,9,20,21,28,29] and, specifically, Tra Vinh Province was scrupulously undertaken [31,32,33]. In Binh Phu Commune, located furthest north, freshwater agriculture was dominant, and farmers mainly practiced intensive rice cropping, with some vegetable and fruit production. In My Hoa and Vinh Kim communes, further south, in addition to double rice and vegetable cultivation, farmers practiced a mixed rice and shrimp system, with either extensive or intensive shrimp farming. Lastly, the southernmost communes of My Long Nam, Ngu Lac, and Don Xuan were dominated by coastal mangroves and shrimp farming, both extensive and intensive. Most of the farms included in this study were located inside the sluicegates. On farms situated outside the protected area, extensive natural aquaculture or intensive shrimp farming was predominant. In each studied commune, farmers on dune ridges made use of rainwater to grow rice in the rainy season while extracting groundwater for dry-season vegetables.

For the selection of interviewees, we adopted the snowball method, commencing with subjects recommended by local agricultural authorities [37]. The number of farmers interviewed in each commune was determined by the point of information saturation, ensuring a comprehensive and rich dataset. To glean insights into the history of land-use practices and the drivers that steer decisions towards changing land uses, we deliberately sought out farmers with extensive farming experience, serving as the primary labor force for their household farms. The choice of semi-structured interviews was judiciously made for this research due to their capacity to encourage interviewees to elaborate, explore novel lines of inquiry, and express their perspectives in their own terms. This approach granted us the latitude to delve into the unique context underpinning each household’s land-use transition, thus fostering a profound comprehension of farmers’ viewpoints and attitudes. However, to conduct these interviews effectively, interviewers need to establish rapport, probe for detailed responses, and allow participants to express their thoughts fully. This process can be time-consuming and necessitates skilled interviewers. As a result, it is often employed with smaller, purposefully selected samples rather than larger, randomly chosen ones.

Initially, our inquiry revolved around historical farming practices, commencing from when the interviewed farmer acquired their property. For each distinct farm, we mapped out an individual trajectory of transitions, delineating significant moments of alteration [38]. Our interviewing approach was semi-structured and focused on discerning the internal and external factors that wielded influence over decisions made on the farms. The individuals being interviewed were regularly encouraged to provide more in-depth elucidations, prompted by queries delving into the mechanics of “why” and “how” certain elements impacted their choices during specific instances of change.

Initial interviews were conducted in January 2016. After preliminary data analysis, we revisited the communes in April 2016 for second interviews with the same farmers or to interview additional farmers considered likely to have different land-use change pathways. Eighteen households were interviewed twice.

We analyzed the interview transcripts to trace the historical development of land use on the farms alongside events in the households and communities. Subsequently, we generated a detailed list of drivers from the empirical data by applying grounded theory with color coding [27,39] (Appendix A, Table A1. presents the list). Grounded theory is an inductive method for incorporating data from interview transcripts into an existing theory or framework. The data acquisition and analysis follow a cyclical process that uses early data analysis to shape the ongoing data collection. Concepts and linkages between different elements in the text are identified by assigning codes through the interview transcript. Color coding is a visual way based on colors to distinguish codes in texts with different concepts or categories. We identified key drivers of land-use change decisions by the frequency they were mentioned by the interviewed farmers [26].

3. Results

3.1. Pathways of Land-Use Change in the Salinity Intrusion Zone

Figure 2 presents pathways of land-use change for the 32 farms studied in the salinity intrusion zone of Tra Vinh Province. Over the 40 years, farmers reported practicing eight farming systems: (i) single rice cropping, (ii) double rice cropping, (iii) triple rice cropping, (iv) a vegetable or fruit monoculture, (v) mixed farming with rice and vegetables, (vi) mixed farming with rice and shrimp, (vii) extensive shrimp farming, and (viii) intensive shrimp farming. Though land uses in the study area were diverse, their geographical expanse was limited by the salinity of the available water resources.

Before the construction of canals to supply freshwater, natural farming systems depended on monsoon rainwater, with saline water intruding in the dry season. Most of the interviewed farmers began their farming careers with a single rainfed rice crop annually. Farmers on the dune ridges could take advantage of freshwater lenses to grow vegetables. Later, they followed different land-use paths, depending on the characteristics of the water at their locations.

Double rice cropping made its appearance before the 1980s, thanks to the advent of high-yielding varieties, but the change from single to double rice cropping mostly happened after 1986, facilitated by decentralization policies. Triple rice cropping was established earliest in Binh Phu (Figure 2, in red) in 1992. Farmers in Ngu Lac and Don Xuan communes (Figure 2, in pink and orange) began triple rice cropping in 2010, after the completion of the South Mang Thit infrastructure intended to secure a year-round freshwater supply.

In My Hoa, Vinh Kim, Ngu Lac, and Don Xuan, most farmers began practicing mixed farming with double rice cropping and vegetables after 1988. Thanks to the specific hydrology of the dune ridges, farmers there were able to grow one or two rice crops in the rainy season and then cultivate vegetables in the dry season using pumped groundwater.

While the South Mang Thit infrastructure was under construction, the canals and natural flows were blocked, severely limiting freshwater supplies over several years and causing a rice production crisis. From 2000 to 2002, farmers in My Hoa and Vinh Kim had to stop freshwater farming altogether. In an attempt to adapt, they reverted to natural aquaculture and the rice–shrimp farming system. One interviewed farmer in Binh Phu had to give up tangerine growing from 2003 to 2010 because sluicegate construction caused a sudden drop in the freshwater level. The farm was re-established after 2010, at which time the water supply and household economic situation had improved (Figure 2, in red).

In the southernmost communes, salinity intrusion was pronounced. Farmers here shifted first from single to double rice cropping and then to raising wild shrimp species in the rice fields. This mixed rice–shrimp farming system, classified here as brackish aquaculture, appeared around 2000. Meanwhile, other farmers converted rice fields into extensive shrimp ponds. However, many of these farmers went back to rice farming after experiencing failure with shrimp. For example, farmers in Ngu Lac (Figure 2, in pink) switched from double rice to extensive shrimp in 2002–2003. However, due to the failure of shrimp production, they went back to double rice cropping with vegetable cultivation in 2006–2007. They also reported stopping farming for several years because shrimp farming had degraded their land. Two interviewed farmers tried shrimp farming again in 2013, while others chose triple rice cropping. Farmers who succeeded in brackish aquaculture typically switched from wild shrimp species to breeding stocks and gradually moved from extensive to intensive shrimp farming. Intensive shrimp requires saline water and high-voltage electricity to run oxygen generation machines. This was witnessed in My Long Nam in 2004 (Figure 2, in purple), followed by My Hoa and Vinh Kim (Figure 2, in green and blue). My Long Nam is a special case where farmers switched directly from single or double rice cropping to intensive shrimp farming thanks to support from the provincial government.

In Don Xuan (Figure 2, in orange), one of our interviewed farmers operated outside the embankment and sluicegates. This farmer reported cultivating one rainfed rice crop annually but with very low production due to the salinity of the water. His family, therefore, tried extensive aquaculture in 2000, catching and raising wild fish, shrimp, and crabs in a shallow pond. Then, they raised brackish shrimp and tried mixed rice and shrimp farming in 2005. More recently, they built an embankment around their ponds for intensive shrimp farming.

When the South Mang Thit infrastructure was complete, salinity intrusion was expected to be under control. However, climate change and sea-level rise have put the system under pressure. The infrastructure was designed with the expectation that freshwater supplies would be sufficient to convert the whole protected area into a permanent freshwater zone. However, freshwater inflows have diminished due to severe droughts, among other factors, and salinity intrusion has been more pronounced than expected. Intensive rice farming has, thus, become increasingly restricted to the northern region and along the inland river banks. As a result, farmers in Binh Phu said they were considering switching back to double rice cropping in 2016. Farmers in the year-round freshwater zone reported reserving a field or garden area to grow vegetables and fruit trees. Meanwhile, intensive shrimp farming gradually extended inland.

3.2. Multi-Scale Drivers of Land-Use Change

Constrained by the available water resources, the farmers in our study area embarked on one of three land-use change pathways. The first is the intensification of freshwater agriculture, in which farmers sought to maximize their cropping seasons by practicing triple rice or combining a single rice season with two seasons of vegetables. The second is brackish-water farming, such as a shift from double rice to a mixed rice–aquaculture farming system (rice in the rainy season and aquaculture in the dry season). The third pathway is the intensification of saline water aquaculture, represented by a shift toward more intensive shrimp farming and higher stocking densities. Farmers reported that their decisions regarding land-use changes were influenced by various drivers at the national, regional, and household scales. Analyzing the interview transcripts, we identified 41 drivers grouped into 13 categories. These included three policy drivers at the national scale, 23 regional-scale drivers, and 15 household-scale drivers (see Appendix A, Table A1).

Figure 3 depicts the frequency with which farmers mentioned each of the multi-scale drivers as affecting their land-use decisions. Drivers mentioned by more farmers were considered to have a more active or significant role in farmers’ decision-making processes.

Very few farmers recognized national policies as playing an influential role in their land-use change decisions (top Figure 3). Many did note the importance of household financial assets, particularly in shifts towards higher yielding or higher value products. More than 10% of farmers said that financial assets were an influential driver of their decisions. As vegetables, fruit, and shrimp require higher investment and are riskier than rice, household financial assets were especially important to enable farmers to shift to these enterprises. Indeed, financial assets were mentioned as important by 14–20% of farmers who reported diversification to vegetable cultivation and extensive or intensive shrimp. Among those practicing a rice monoculture, 11% mentioned financial assets as an important influence in land-use decisions (Figure 3).

Early in the study period, a single rainfed rice crop was the only feasible farming option in much of the salinity intrusion zone due to limitations imposed by the acid sulfate soils and freshwater shortages in the dry season. However, this provided a meager livelihood, as rice productivity was low, and lands were often left uncultivated in the dry season. The shift to double rice cropping was facilitated by two main factors: mobilization by local authorities (institutions), mentioned by 16% of the relevant farmers, and development of the canal system (infrastructure), mentioned by 14% of relevant farmers (Figure 3). Farmers in Binh Phu said that in the 1980s, local governments mobilized them to join cooperative groups to dig canals and ditches to bring in freshwater from the Mang Thit River in northern Tra Vinh Province. Expansion of the canal system not only increased freshwater inflows to remote and coastal areas but also prolonged the months of freshwater availability, as the canals retained rainwater, ensuring its availability for soil flushing and irrigation. In addition, local extension agents introduced high-yielding rice varieties, which gradually replaced the rainfed varieties. Thus, local authorities played a significant role in advancing the switch to double rice cropping in the region. Later, as natural assets improved and farmers saw the success of double rice cropping in the community, they followed the trend.

Although local authorities also played a role in mobilizing farmers to try triple rice cropping, we found the significance of institutions (8%) in this land-use change to be less than that of the socio-economic context (17%), markets (16%), infrastructure (13%), and natural assets (13%) (Figure 3). Most farmers said their desire to shift to triple rice cropping stemmed from witnessing successes with this farming system within their community. The development of rice markets, trader networks, and expanded agricultural services were also mentioned as supporting the intensification of rice farming. Last but not least, rice-based agricultural intensification in the coastal estuarine zone would have been unachievable without large-scale hydraulic infrastructure, especially the South Mang Thit project, as freshwater canal construction gradually reached remote areas, and sluicegates were installed to control salinity intrusion.

A mixed rice and vegetable farming system was easily embraced on the dune ridges because a much smaller quantity of water was needed for irrigating vegetables compared to rice cropping. Farmers with these lands had access to groundwater, which allowed them to maximize their cropping seasons, cultivating vegetables during the freshwater-scarce dry season. Overall, the switch from rice farming only to a mixed rice and vegetable system was influenced by four main drivers: suitable natural assets (20%), availability of financial assets (15%), appropriate technologies (13%), and favorable markets (11%) (Figure 3). Farmers frequently mentioned suitable natural assets (farm conditions) as a prerequisite for this land-use change. Lands had to be at a high enough elevation with a sandy soil type to prevent vegetables from waterlogging. Furthermore, most of these farmers irrigated using rainwater stored in backyard ponds or nearby canals. On dune ridges, where vegetable growing was common, irrigation was highly dependent on groundwater extraction. Since market prices for vegetables were higher than for rice, farmers believed they could earn a greater profit from vegetables, even though vegetable farming required a larger investment than rice. Farmers used mechanization, agrochemicals, and fertilizers to optimize production and ensure a successful season even under unfavorable conditions. They also participated in training workshops organized by agricultural officers and companies to access new techniques and financial support.

The South Mang Thit project had a substantial impact on maintaining freshwater conditions across much of Tra Vinh Province. However, the sluicegate network could not entirely prevent salinity intrusion and provide advantageous water-use conditions for all farmers. According to one interviewed farmer in My Hoa, a water crisis forced him and his neighbors to switch from double rice cropping to a mixed rice and shrimp farming system. These farmers, at first, were very satisfied with double rice cropping and the freshwater supply system. However, in the late 1990s, the government of Tra Vinh Province started to build a large sluicegate system along the main rivers to protect the province from salinity intrusion. During construction, natural flows were blocked, resulting in severe shortages of freshwater and huge damage to rice production in eight communes along the Co Chien River. This extreme hardship lasted for three years (1999–2001). During that period, farmers had to leave their lands uncultivated, with many moving to other regions to work as laborers. Eventually, tension grew to the point where farmers in the affected communes broke the infrastructure to allow water to enter their fields. To resolve the crisis, the provincial and local authorities reached a compromise with the farmers, adjusting the sluicegate operation (2001–2002). However, double rice cropping hardly returned to the region. Water quality remained poor, and the acid sulfate soils were unconducive to farming. Some areas became brackish due to the leakage of saline water. Changes in natural assets here forced farmers and local authorities to seek an alternative farming system. Following the example of farmers in the saline zone, such as Duyen Hai District, they began to raise shrimp in the rice fields. Some dug ditches around their rice paddies to raise shrimp. Others raised freshwater shrimp in the rice fields in the rainy season, while in the dry season, they let brackish water inundate their fields to raise wild shrimp and fish. Influential land-use change drivers for these farmers were financial assets (18%), natural assets (14%), and the socio-economic context (11%). Infrastructure (9%) and institutions (9%) were two other key drivers of this pathway of land-use change (Figure 3).

In a less contentious context, farmers sometimes chose extensive shrimp farming as an intermediate land use between freshwater agriculture and saline aquaculture. For this land-use change pathway, two groups of drivers were distinguished, depending on farm locations and their corresponding agro-hydrological conditions. Rice farming was unproductive on farms outside the embankment and sluicegates due to the dominant role of tidal regimes and saline water from the sea. In this high-salinity zone, both farmers and local authorities actively sought alternative farming options to improve household livelihoods. A farmer in Don Xuan observed that he and his neighbors were able to switch from a single rice crop to extensive shrimp farming thanks to the encouragement and technical and financial support of local authorities. The multi-scale drivers considered most influential in this change were natural assets (10%), institutions (14%), and financial assets (14%) (Figure 3). Within the protection of the embankments and sluicegates, extensive shrimp farming was more of a choice made by farmers. For instance, farmers in Ngu Lac were tempted when they saw the profits others were earning from shrimp farming and the market potential. Therefore, they deepened their rice fields slightly and pumped saline water into the fields—though this was strictly prohibited by local authorities as it risked undermining the quality of neighboring fields. The main multi-scale drivers in this land-use change were socio-economic context (16%), financial assets (14%), technologies (12%), and the market (10%) (Figure 3).

Change pathways involving shrimp production were diverse. Depending on their natural and financial assets, some farmers reverted back to rice or chose to upgrade to more intensive shrimp farming. We found intensive shrimp farming being increasingly practiced inside the South Mang Thit infrastructure, despite efforts of many farmers and communes to maintain freshwater agriculture within the protected area. Most of the interviewed shrimp farmers admitted they had switched to intensive shrimp farming in the hope of earning much greater profits than in other land uses. Yet, there were also farmers who said they had moved to intensive shrimp farming reluctantly after an unsuccessful extensive shrimp operation or mixed rice–shrimp enterprise. Moreover, salinized and polluted water from surrounding shrimp farms made some lands unsuitable for rice farming. However, intensive shrimp required good infrastructure, particularly a strong and stable electricity source, because pumps and oxygen generators were a prerequisite for dense stocking. In addition, farmers had to obtain loans from banks, supply agents, or even the black market to purchase breeding stock, medicines, and feeds. Meanwhile, the shrimp market was precarious and insecure. Though shrimp could be very profitable, failure due to disease was a constant worry. Farmers mentioned the heavy burden of the accumulated debts to suppliers, saying they hoped only to be solvent, notwithstanding the potentially high returns from intensive shrimp cultivation. The main drivers in this land-use change path can be summarized as financial assets (20%), infrastructure (12%), socio-economic context (12%), natural assets (10%), and technologies (9%) (Figure 3).

In My Long Nam specifically, where local authorities had developed a plan for intensive shrimp farming after the water crisis, our study also identified a context driver in that institutions played an important role in achieving adjustments in the operation of the sluicegates (12%).

3.3. Spatial and Temporal Aspects of Drivers

This section examines the spatial and temporal aspects of land-use change drivers. Obviously, geography has had a substantial impact on the hydrological conditions and land uses in the studied communes. For instance, the more northern farms and those within the South Mang Thit infrastructure had easier access to freshwater via canals and were less exposed to salinity intrusion. Indeed, in the freshwater zone, intensive rice farming and fruit and vegetable cultivation were still favored, while extensive and intensive shrimp farming was dominant in the more saline areas. Figure 4 depicts the frequency with which the different drivers were mentioned in the studied communes. We found that land-use change drivers indeed varied according to agro-hydrological settings, even at the small-scale studied sites.

The transition from single rice cropping or a rice–shrimp farming system—both of which can be considered traditional land uses—to double rice cropping occurred in all six studied communes, although the change differed in space and time (see Figure 2). In general, improved natural assets, thanks to infrastructure development, played a key role in enabling irrigation and intensive rice farming. However, the interviews in Binh Phu underlined divergence in this driver. All of the farmers there mentioned national policies and local institutions as being the greatest forces for land-use change at the time (Figure 4). This is reasonable because these farmers started double rice cropping in the early 1980s when the centralized agricultural production policy was dominant. Binh Phu also had the advantage of easier access to freshwater since it is situated in the north, closer to the Mang Thit River (see Figure 1). Favorable natural assets and socio-economic conditions enabled farmers in Binh Phu to switch to triple rice cropping straightforwardly. The farmers interviewed here also mentioned the important role of the ready availability of agricultural services and rice markets in the change to intensive rice farming. Later, triple rice cropping emerged in the southern region, too, such as in Ngu Lac and Don Xuan, mainly in association with the southward expansion of freshwater canals and the sluicegate system.

For the switch from a rice mono-crop to a mixed rice and vegetable farming system, natural assets were key, particularly elevated lands with good drainage capacity. These features were found on the coastal dune ridges in the more southern communes (My Long Nam, Ngu Lac, and Don Xuan). However, vegetable cultivation was not considered possible without technical interventions and fertilization. Finally, the high market price of vegetables helped to reinforce household incomes, boosting financial assets.

The move to a mixed rice and shrimp farming system in My Hoa and Vinh Kim was somewhat different. Farmers in My Hoa gave up double rice cropping because the construction of sluicegates degraded their freshwater resources and induced seasonal brackish conditions. Thus, the start of shrimp farming in My Hoa was more an attempt to solve a problem caused by the hydraulic infrastructure than a move desired by farmers. In contrast, farmers in Vinh Kim were motivated by their neighbors’ success with shrimp. They made adaptations to retain brackish water in their rice fields to raise shrimp in the dry season with high hopes that a successful harvest would increase their financial assets.

We interviewed extensive shrimp farmers in different geographical settings. The farms in Ngu Lac were protected by dikes and sluicegates. Decisions to try extensive shrimp farming here were driven by economic trends in the community, the attractive shrimp markets, and the available natural and financial assets. Some farms in Don Xuan were located outside the protection of the hydraulic infrastructure. Key drivers of extensive shrimp farming here were the search for livelihood solutions for the open water environment, support from local authorities, and the availability of technologies.

Decisions to shift to intensive shrimp production by farmers in My Hoa, Vinh Kim, and My Long Nam were influenced by similar drivers, especially a strong and stable electricity source (infrastructure), availability of funds to invest (financial assets), and the desire to earn as great a profit as neighbors (socio-economic context). Shrimp farmers in Don Xuan emphasized the role of socio-economic context and financial assets, though they did not have full access to a stable electricity source and the protection of dikes and sluicegates.

3.4. Cross-Case Comparison

Comparing the land-use change pathways found among farms in our salinity intrusion case with those from our earlier case study of the highly flooded upper delta zone [26] (Appendix B 

Table A4. A cross-case comparison of land-use change pathways and driving factors in An Giang and Tra Vinh Provinces.

Case Study	Highly Flooded Zone—An Giang Province *	Salinity Intrusion Zone—Tra Vinh Province
Hydrological characteristics	The highly flooded zone consists of the Long Xuyen Quadrangle in the west (An Giang province) and the Plain of Reeds in the east (Dong Thap province). Between them lies a higher altitude freshwater alluvial zone with favorable agricultural conditions due to year-round freshwater availability and good soils (also belonging to An Giang province). In the highly flooded zone, agriculture used to be limited by soil conditions, which had high acid sulfate levels in the dry season (November to April) and by the rising and falling of floodwaters in the rainy season (May to October). These floodwaters start to rise slowly, by 10–15 cm per day, in July or August, and then peak at 3–4 m in late September or early October. The annual flooding regime and soil conditions are managed by a system of drainage canals and high dikes built by the government. In An Giang province, the topography of the eastern region is completely flat, while there is a small mountainous area with forest in the western region. The eastern region had a higher percentage of area protected by high dikes compared to the western region.	The salinity intrusion zone in Tra Vinh is influenced by the Mekong River system’s waters and the East Sea’s tidal regime. Constraints include salinity intrusion and acid sulfate soils, especially in the dry season. Freshwater sources in the province include the Bassac River to the west, the Co Chien River to the east (a tributary of the Mekong River), and the Mang Thit River to the north. To the south, the province is bordered by the East Sea, from which saltwater easily intrudes inland through the river mouths. Lower dry season discharge from the Bassac and Co Chien rivers enables saltwater to penetrate north, also affecting the western and eastern parts of the province. The province benefits from the South Mang Thit salinity prevention infrastructure system, involving a system of canals, dikes, and sluicegates overseen by the central government. Farms on dune ridges can access freshwater lenses (groundwater) for irrigation, making these traditionally ‘poor’ lands attractive in the saline dry-season environment.
Land uses	Freshwater agriculture predominates in the studied area, including rice farming systems, vegetable cultivation, and a limited number of fruit tree gardens. Intensive rice cultivation (triple rice cropping) has been facilitated by government efforts to control the flood regime and improve soil quality.	Farmers have the flexibility to practice various farming systems. In the rainy season, farmers optimize rainwater for rice cropping. In the dry season, farmers shift to aquaculture due to freshwater scarcity. Canals and sluicegates have enhanced freshwater supply, prevented salinity intrusion, and enabled freshwater agriculture, particularly double and triple rice cropping. However, ongoing issues like drought and changing rainfall patterns continue to threaten freshwater agriculture and exacerbate salinity intrusion.
Land-use change pathways from 1975 to 2016	The prevalent trend is a uniform, stepwise rice intensification, with a growing interest in diversifying into vegetable farming. Five main systems were identified: single rice cropping, double rice cropping, triple rice cropping, mixed farming with rice and vegetables, and vegetable monoculture. Farmers initially practiced single rainfed rice cropping due to the effect of acid sulfate soils and annual flooding. In the 1980s, the government-led acid sulfate land reclamation and introduction of high-yield rice varieties facilitated double rice cropping. With the Doi Moi political reforms, the land was returned to previous owners, leading to a resurgence of double rice cropping. Triple rice cropping became viable in the early 2000s with the construction of high dikes that completely protect the area from annual floods. Few farmers transitioned to vegetable monoculture due to challenges posed by annual floods, high dikes, and triple rice dominance. Mixed farming, involving rice and vegetables, emerged outside high dike enclosures or on elevated beds, allowing for rotation between rainy-season rice and dry-season vegetables. Western communes adopted triple rice cropping later than eastern ones, but in general, changes in land use occurred concurrently within communes due to specific water conditions or historical events.	A more diverse shift towards saline-based shrimp intensification in a constrained salinity intrusion zone: (1) farmers primarily adopted freshwater agricultural intensification through triple rice cropping or mixed rice–vegetable systems, (2) shifted to brackish-water farming like mixed rice–aquaculture, or (3) embraced intensive saltwater aquaculture, particularly shrimp farming. Eight distinct farming systems were identified: single rice cropping, double rice cropping, triple rice cropping, various monocultures (vegetables, fruits), mixed farming (rice with vegetables or shrimp), and both extensive and intensive shrimp farming. Initially, most farmers practiced single rice cropping, using rainwater. Some farmers in the dune ridges were able to grow vegetables in the dry season using groundwater. The shifts from single to double cropping mostly happened after 1986, facilitated by the Doi Moi political reforms and canal expansion. Triple rice cropping was initiated in the 1990s following the South Mang Thit infrastructure’s completion for year-round freshwater supply. Salinity challenges led to shifts from rice to mixed shrimp–rice farming, while brackish aquaculture involving wild shrimp species appeared around 2000 in southern communes. The study highlights diverse transitions driven by water characteristics.
Key driving factors The percentage value presents the frequency with which farmers mentioned each of the driving factors	Farmers emphasized the desire for higher agricultural profits as a consistent motivation (12.2%, 12.2%, 15.6%, and 21.7%). Nevertheless, the study showed that for those following the rice intensification path, regional-scale drivers had a more prominent influence than household finances. The transition from single to double rice cropping was primarily propelled by local institutions (16.6%) and community opinion, in line with food security policies (7.4%). Socio-economic context (11.8%) and infrastructure development (8.3%) also played vital roles in this transition. Neighbors’ success with double cropping encouraged others to follow suit. Water availability, influenced by geographic proximity to rivers and canals (7.9%), further stimulated this shift. For the progression from double to triple rice cropping, market-related factors (21%) gained significance alongside institutions (14%), infrastructures (10%), and socio-economic context (10.9%). High dikes were pivotal in enabling triple cropping by preventing floods. Market demands and potential profits became substantial drivers. Mixed farming with rice and vegetables, though occupying a small area, was driven by market incentives (25.6%). Other farm-level drivers for a switch to the mixed farming system were financial assets (12.2%), human assets (5.6%), and the physical and natural assets of the farm (10% and 12.2%). Our analysis also suggests that only farmers with strong financial (21.7%) and labor capacity (17.4%) could embark on a specialization in vegetables. Farmers saw higher vegetable prices as a motivator, with successful vegetable seasons promising greater income than intensive rice farming (26.1%). However, vegetable cultivation posed challenges due to its capital-intensive nature, perishability, labor demands, and unstable markets.	Institutional mobilization (15.8%) and canal development (13.9%) played crucial roles in promoting double rice cropping, with local authorities fostering cooperative efforts and providing high-yielding rice varieties. Triple rice cropping’s adoption was more influenced by socio-economic context (16.7%), markets (15.5%), infrastructure (13.1%), and natural assets (13.1%) within the community. Hydraulic projects like the South Mang Thit facilitated rice intensification by improving freshwater availability and extending canal networks. Shifts to mixed rice–vegetable farming were limited to specific environmental conditions (20%), where freshwater lenses allowed for irrigation. As vegetable farming required higher investments but offered better returns, financial assets (15%), appropriate technologies (12.5%), and favorable markets (10.8%) played a significant role. However, increasing drought and salinity intrusion are putting stress on this system. Shrimp conversion varied in terms of intensity and duration, from intermittent rice–shrimp farming to extensive shrimp farming, depending on local agro-hydrological conditions and socio-economic context. The construction of sluicegate systems initially disrupted freshwater supplies and led to rice crises, prompting shifts to brackish aquaculture, such as mixed rice–shrimp systems. Influential land-use change drivers for these farmers were financial assets (18%), natural assets (14%), and the socio-economic context (11%). Infrastructure (8.9%) and institutions (8.9%) were two other key drivers of this pathway of land-use change. Outside the embankment and sluicegates, the main drivers in extensive shrimp farming were natural assets (10%), institutions (14%), and financial assets (14%). Within the protection of the embankments and sluicegates, extensive shrimp farming was more of a choice made by farmers. The main drivers in this land-use change were socio-economic context (16%), financial assets (14%), technologies (12%), and the market (10%) Intensive shrimp farming arose due to potential high profits but also posed risks, including disease outbreaks and debt accumulation. Thus, the ability to shift to intensive shrimp farming was highly dependent on financial investment and carried associated risks. The main drivers in this land-use change path can be summarized as financial assets (19.6%), infrastructure (12.4%), socio-economic context (12.4%), natural assets (10.3%), and technologies (9.3%)
Highlights	Land-use changes in An Giang Province were uniform due to collective community actions, such as building dikes and canals to control floodwaters. Additionally, high-yielding rice varieties, supported by strong export markets, influenced the shift towards rice intensification. Biophysical constraints in polder-based paddy rice areas limited diversification efforts.	Tra Vinh Province witnessed more diverse land-use change pathways over time, influenced by policy and infrastructure developments. While these developments aimed at rice intensification, they failed due to factors such as drought, salinity, and the lucrative shrimp market.

* Sources: Le, T.N.; Bregt, A.K.; van Halsema, G.E.; Hellegers, P.J.G.J.; Ngo, T.T.T. Multi-Scale Drivers of Land-Use Changes at Farm Level I: Conceptual Framework and Application in the Highly Flooded Zone of the Vietnamese Mekong Delta. Land 2023, 12, 1273.

), two primary pathways of land-use change emerge. The first is a uniform, stepwise pathway of rice intensification, most evident in the floodplains of An Giang Province, with a nascent diversification drive toward vegetables. The second is the markedly more divergent pathway towards saline-based shrimp intensification, as evident in the coastal estuary zone of Tra Vinh Province. On a regional scale, the land-use change pathways found in both cases demonstrate temporal and spatial diversity, elaborated in more detail below.

Land-use changes were rather uniform in An Giang Province due to two factors: the biophysical environment and community-level collective action. Particularly, collective action was needed to construct the low and high dikes required to control the entry of floodwaters into the rice fields and to adjust water levels for drainage and irrigation. Such an infrastructure system could not be constructed or managed by an individual farm on its own. Attesting to this, interviewed farmers provided accounts of collective community action mobilized by local authorities to reclaim floodplains, dig canals, and construct low dikes, marking the transition from single to double rice cropping. In the shift from double to triple rice cropping, community action played out differently. Community agreement (60%) needed to be reached and formalized before high dike construction could begin. The process of gaining community agreement was sometimes relatively swift, but it could also be prolonged. In one community, it took some 10 years to achieve [26]. This determined when the switch from double to triple rice cropping occurred in a particular locality, highlighting the importance of institutional and community factors in shaping this land-use change.

Uniformity was also imposed by the biophysical constraints to agricultural diversification in the polder-based paddy rice environment. Here, high costs and risks impeded other land uses, such as the cultivation of vegetables, fruits, or other dryland crops. These alternative products were also highly perishable and not served by a strongly developed market and services network in the study area, further increasing the financial risk of switching from rice to, for example, vegetable cultivation. Combined with the growing shortage of labor, these factors explain the uniformity of the rice intensification pathway, as we observed only one conversion to an entirely vegetable operation in our case studies.

While land-use change in An Giang Province was uniform, spatial and temporal diversity was found. This depended on the sequence and timing of land reclamation and when the canals and dikes were constructed that enabled the floodplains to be transformed into the ‘rice bowl’ of Vietnam. This process advanced from east to west. Land-use change neatly followed a spatiotemporal pattern wherein the shifts from single to double and from double to triple rice cropping were determined by the place and time of floodplain reclamation, serviced by low dikes and canals, and the transition from low to high dikes for further control over agro-hydrological conditions. As such, infrastructural developments were instrumental in enabling and shaping this land-use change pattern over time and space.

Land-use changes were also shaped and reinforced by the particular combination of drivers experienced by farmers. The introduction of high-yielding rice varieties was a key technological innovation enabling rice intensification. Their widespread application throughout the VMD, particularly in the upper delta floodplains, was reinforced by the establishment of a strong and dedicated export market and agricultural support services for rice, propelled by national government policy in line with its goal to achieve rice-based food security. The importance of market factors increased over time and with greater intensification. Markets shaped land-use change in An Giang Province in three ways: (i) by providing a secure and low-risk outlet for production, (ii) by strengthening the intensification pathway as a strategy for income generation, and (iii) by stimulating consolidation (scaling up) and mechanization of rice farming to secure greater income from this low-value staple. As such, we found that toward the end of the study period, farmers were migrating from eastern to western An Giang Province to secure larger rice land holdings. At the same time, the outmigration of young people in search of better-paying urban jobs introduced labor shortages as a constraint, and the advancing age of the region’s rice farmers seems likely to propel further rice farm consolidation and mechanization in the future.

In Tra Vinh Province, the land-use change pathways identified between 1975 and 2016 (Figure 2) were more divergent, encompassing both freshwater and saline cropping systems and a high dynamic of change over time. Here, policy and infrastructure developments, especially the construction of the South Mang Thit project and sluicegate system, were geared towards a rice intensification pattern similar to that found in An Giang Province. However, this did not materialize. The infrastructure system was designed to ensure freshwater availability year-round for an agro-hydrological environment conducive to rice and fruits. However, it failed to provide this uniformly across the command area as the result of interacting push and pull factors: (i) the ever-increasing impact of drought and salinity intrusion, which has fueled shortages in freshwater supply, and (ii) the emergence of a strong international market for shrimp after the shrimp boom of the 1990s. Regarding the first, deteriorating growing conditions in the dry season, with freshwater supplies falling short, polluted stagnate water behind closed salinity control gates and salinity penetrating beyond sluicegates into formerly freshwater zones pushed farmers towards saline aquaculture, particularly as they experienced increasing failure in the second and third rice crop or other dry-season crops. The pull factor of an accessible market for high-value shrimp has attracted farmers who are interested in potentially higher incomes. Some farmers have actively sought to shift from rice to shrimp farming, for example, opening sluicegates or pumping saline water into their fields to enable shrimp to be cultivated among rice fields. Unlike the accounts of vegetable conversion in An Giang Province, the shift to shrimp farming in Tra Vinh Province can be pursued individually. It does, however, affect neighboring rice fields through salinity leakage and, thus, often forebodes a wider shift to a more saline environment. The ‘strictness’ of community institutions’ adherence to formal land-use plans is indicative of the swiftness and degree of the saline-oriented land-use changes that are allowed to occur.

The spatial and temporal pattern of shrimp conversion has been determined by market forces (the post-1990s boom), local hydrology (distance from a freshwater source or the coast), and the severity of the drought year (and its associated salinity intrusion into the river arms, which has steadily increased post-2016). The salinity gradient (from saline to fresh) runs from south to north and from the coastal edges land inwards. With the South Mang Thit project, the government attempted to introduce a hard boundary between permanent saline-intensive shrimp in the south and a freshwater zone in the command area of the project. At the outer and southern edges, however, this boundary has been increasingly undermined by the combined impacts of the push and pull factors described above. The communes at the edges of the South Mang Thit project (Don Xuan, Ngu Lac, My Long Nam, My Hoa, and Vinh Kim) have progressively shifted toward shrimp cultivation as growing conditions on their lands further deteriorate, and the effects of pioneer shrimp converters favor a brackish environment. As more farmers convert to shrimp, the quest for good quality saline water increases, leading to gradually increasing salinization of the local agro-hydrology. This has been particularly discernible in the southeastern part of the studied area, where salinity levels have increased and the South Mang Thit salinity barrier has been moved inland, effectively opening up parts of the system to salinization and aquaculture.

The shift to shrimp cultivation in Tra Vinh Province has been neither uniform nor unilateral. The intensity and duration of shrimp farming have varied, with interviewees reporting intermittent rice–shrimp farming, extensive shrimp farming, and year-round intensive shrimp farming (typically with three harvests annually). Overall, the shift from a freshwater rice monoculture to saline shrimp cultivation has been moderated by local agro-hydrological conditions and the socio-economic context at the household scale. Due to the push factors of drought, freshwater shortages, and salinity intrusion, dry-season rice cultivation is increasingly undermined, with severe yield losses or the need to leave lands fallow. The natural and easiest conversion, then, is a shift to intermittent rice–shrimp farming, in which a rainfed rice crop is followed by dry-season shrimp cultivation, facilitated by the active intake of saline water in the dry season. As shrimp incomes rise and salinity intrusion becomes more severe and prolonged (both due to natural causes and active intake), expansion of shrimp cultivation and then to extensive shrimp farming seems likely (as clearly observed in My Hoa).

The shift to and maintenance of intensive shrimp cultivation, however, is highly dependent on financial investment and carries high associated risks. The conversion to intensive shrimp is capital- and technology-intensive, while the risk of disease-induced crop failure (associated with high-density stocking and lack of access to wastewater disposal and clean saline water intake) makes intensive shrimp farmers susceptible to indebtedness and bankruptcy. Bankruptcy and financial burdens may, thus, induce intensive shrimp farmers back into extensive shrimp or rice–shrimp farming. Likewise, access to capital determines whether farmers follow a stepwise intensification path or opt for a more radical shift from double rice cropping directly to intensive shrimp farming. Financial assets are, thus, highly associated with the ability to shift to shrimp cultivation and the choice of intensification pathway a farmer may take (see Figure 3 and Figure 4).

The switch to a mixed rice–vegetable farming system is limited in scope, being confined to the environmental conditions of the dune ridges. Here, rains replenish freshwater lenses, from which farmers may draw water to irrigate vegetables during the dry season. This natural asset is the key factor enabling diversification into vegetables in an otherwise drought and saline-prone environment. The availability of freshwater lenses, therefore, determines the spatial distribution of vegetable cultivation in Tra Vinh Province. However, with drought and salinity intrusion being increasingly severe and of longer duration (starting earlier and lasting longer) since 2016, this system is increasingly under stress. Freshwater lens capacity, too, is diminishing, forcing farmers to reduce their vegetable cultivation to two rather than three or four short-season crops and limiting their options for other crops.

4. Discussions

The application of the multi-scale drivers framework in the two cases brought some generic lessons to the fore. First, the adoption of the historic perspective over the 40-year timeframe revealed the different steps in land-use change and emergent pathways forming distinctive patterns of change [28]. The spatial and temporal distribution of land-use change was marked by infrastructure development to control the agro-hydrological environment [40,41,42,43]. Our findings regarding the spatial distribution of land-use change align with those reported by Le et al. (2018) [28] for the VMD. Our detailed and historic study, however, in addition, demonstrates that present-day land uses are the outcome of particular pathways of land-use change, characterized by defined steps of intensification or diversification. These different pathways and steps may account for the highly dynamic 15% annual average land-use change reported by Le et al. (2018) [28].

Pathways of land-use change and path dependencies [25,38] were shaped by the communal and spatial effects of the specific agro-hydrological conditions created and entered into at each step. This was most pronounced in An Giang Province [26], where the shift from single to double and to triple rice cropping had to be made communally to enable hydraulic infrastructure, such as diked polders, to be built and managed at the appropriate scale. The uniformity of land-use change was further reinforced by the paddy rice environment, in which a shift to vegetables was extremely difficult. In Tra Vinh Province, in contrast, a shift from rice to shrimp farming could be initiated at the individual farm level as long as saline water was proximate, enabling more diversity in the change pathway. Once started, however, salinity affected neighboring rice fields, drawing them into the salinity pathway as well.

Multiple factors, at multiple scales, played a role in driving land-use changes in our case study areas. We, thus, conclude that land-use change cannot be attributed to a single factor but is an outcome of an interaction between or specific configuration of multiple factors [20,21,22,23,24,25,26]. Farmers’ responses to drivers attest to this. However, as no ranking or causal chain analysis was applied in our methodology, no clear-cut and generic configuration of factors can be derived for any of the emergent land-use change pathways. This would require follow-up research in which the pathways are discussed with respondents, as well as the conditions that must be met in order to follow these paths and the factors contributing to these conditions.

The frequency distribution in which the factors were mentioned by respondents was different for each land-use change step (see Figure 3). To a lesser extent, the frequencies differed for each commune (see Figure 4) in both An Giang and Tra Vinh provinces. Though independent causality cannot be attributed to even the most frequently mentioned factors, the ones mentioned most do provide insight into what drivers farmers perceived as most critical in enabling or constraining change at each juncture. Examples of key factors are the institutional requirements for collective reclamation of the floodplains to enable the shift from single to double rice cropping, dependency on access to the natural asset of freshwater lenses for dry-season vegetable cultivation, and access to financial assets to shift to high value and intensive vegetables in An Giang Province [26] or to intensive shrimp farming in Tra Vinh Province. These seem indicative of the difficulty of looking back on and explaining path-dependent changes from a historical perspective.

A land-use change step represents a moment in time in which farmers decide under various influences to switch from one land use to another. That such change may be in part dependent on the changes that took place before it can easily escape attention. To address this more specifically in the future, a two-step analysis approach could be applied: first with a mapping of pathways of land-use change and the multiple factors in play, followed, second, by a joint path and factor dependency analysis of the results conducted, for example, in a focus group discussion.

In this regard, it warrants mentioning that our method of examining the influence of drivers from the farmers’ perspective could have led to bias. Specifically, it may have limited our view of higher-scale effects, as the farmers tended to focus on local and regional factors. National policies were distant and often became noticeable at the farm level only after they had been implemented through projects in the region. To prevent such bias in the future, additional information should be collected from local officials and experts with knowledge of policies at the regional and national levels.

Furthermore, the utilization of semi-structured interviews and transcript analysis imposes demands in terms of time and resources. Consequently, our interview scope was confined to 32 farmers distributed across six communes. Despite this limited sample size, our objective is to ensure the relevance and representativeness of our findings and conclusions within the context of the salinity intrusion zone’s population. Achieving this goal necessitates that the analyst possesses a comprehensive understanding not only of the interview context but also of the geographic, demographic, and cultural setting of the study areas. In light of these considerations, a literature review on land-use change in the salinity intrusion zone was conducted prior to initiating the interviews [3,6,9,20,21,28,29,31,32,33]. The six studied communes were selected with consideration to cover the regional heterogeneity. Leveraging recommendations from local agricultural authorities, we were able to interview farmers representing diverse land-use types present in the six communes. Additionally, nearly half of our participants underwent a second round of interviews with additional inquiries. This approach facilitated both confirmation and in-depth exploration of our interview findings but might constrain the reproducibility of our research in more expansive regions with diverse interview groups. In light of our previous study in An Giang Province, we propose that studies conducted on a broader scale consider alternative methodologies for data collection to mitigate processing time [19,20,21,26]. For instance, the trajectories of land-use changes and the key driving factors can be explored using methods such as an examination of existing literature, interviews with key experts, or in-depth group conversations. Subsequently, the measurement of the impact of these driving factors can be accomplished through the administration of surveys featuring a substantial number of participants. This recommendation aims to enhance the feasibility of research replication and extension across varied geographic regions and interview demographics.

5. Conclusions

Our cross-case comparison revealed that land uses in the coastal estuarine zone were more diverse than those in the upper delta floodplains. This is explained mainly by the geographical setting of the coastal estuary, especially the alternating fresh and saline water conditions and the ability of individual farmers to initiate land-use change on their own. We observed diversification not only in land uses but also in the land-use change pathways. Farmers chose to embark on rice intensification, to integrate rice cropping with vegetables or aquaculture, or to gradually switch to intensive shrimp farming. Limitations imposed by soil and water conditions and the desire to earn greater profits were key drivers of land-use change. However, facilitating each land-use pathway were other drivers at the international, national, regional, or local scale. The freshwater supply canals and the sluicegate system built under the South Mang Thit project played a remarkable role in the agricultural development of the coastal estuary. The project was just one of numerous manifestations of the rice-focused national food security policy. However, the hydraulic interventions caused local environmental degradation, eventually undermining agricultural production and forcing some farmers to seek alternative livelihoods and land uses. Finally, important roles were played by the socio-economic context, particularly trends and pressures from the community and markets.

Our study introduces a systematic and scientific approach to the analysis of land-use history, considering the full set of biophysical and socio-economic drivers of land-use changes at the farm level. Our multi-scale framework enabled us not only to explore the richness of qualitative data through farmers’ narratives but also to quantify the contribution of each driver by the frequency with which it was mentioned by interview respondents. This quantification provides a basis for cross-case comparison between different agro-hydrological settings. The developed framework [26], moreover, proved effective in categorizing drivers of land-use change, as well as in structuring their interactions. This framework, alongside our findings in the upper floodplains and coastal estuarine zone, enriches our understanding of land-use changes and their underlying drivers across diverse agro-hydrological zones in the VMD. By delving into the driving factors of land-use changes as perceived by local farmers in our two case studies, we offer a valuable platform for the evaluation and discussion of factors operating at more localized scales as opposed to the national policy and its subsequent regional water interventions. Additionally, our findings serve as a valuable reference point for similar research endeavors in other regions of the VMD.