Assessing the Effect of Intensive Rice Monoculture on Land Degradation Under the SDG 15.3.1 Framework

Abstract

Rice monoculture systems, often involving double- or triple-cropping cycles annually, require intensive agricultural practices that can lead to land degradation. This study evaluates land degradation within the long-term rice monoculture systems of Nakhon Sawan, Thailand, using the Sustainable Development Goal 15.3.1 framework. By focusing exclusively on persistent rice-growing areas, the study minimized the confounding signals of land-use conversion, allowing for an evaluation of the trajectories driven by combined agricultural management and climatic factors. The assessment integrated land use and land cover (LULC), soil organic carbon (SOC) stocks, and land productivity. Findings indicate that 83% of the original paddy area remained long-term monoculture, with LULC-related degradation limited to 4% of the original paddy cultivation area. While SOC depletion was observed in a few districts, a broader potential carbon accretion trend was identified across the province, likely driven by sustainable post-harvest practices such as stubble retention and organic amendments. Land productivity analysis revealed partial stress only in a few districts. The study demonstrated that long-term rice cultivation did not result in widespread deterioration of soil health on an aggregate provincial scale; however, district-localized degradation hotspots suffering from soil organic carbon depletion and climate-induced productivity stress were identified, demanding targeted regional management.

1. Introduction

Land is an indispensable resource, fundamental to both environmental integrity and human livelihoods, as it underpins global food security and provides a vital foundation for economic development [1,2]. However, this resource is increasingly compromised by land degradation, defined as the persistent decline in biological or economic productivity resulting from the complex interplay between natural processes and anthropogenic pressures [3,4]. This degradation is primarily driven by global climate change, poor land-use management, and the intensifying food demands of a rapidly growing population [1,5].

Land degradation reduces land productivity, thereby posing a significant threat to livelihoods, particularly among rural populations whose incomes depend heavily on agriculture [6,7]. Currently, approximately 20% of the Earth’s land surface, comprising 20% of croplands, 10% of grasslands, and 30% of forests, is classified as degraded. This vast area spans two billion hectares and directly affects the well-being of more than three billion people [1,8].

To address this, the United Nations Convention to Combat Desertification (UNCCD) established the Sustainable Development Goal (SDG) indicator 15.3.1 framework [8,9]. This methodology evaluates the proportion of degraded land through three key sub-indicators: land cover, land productivity, and soil organic carbon (SOC) stocks. While this framework has been successfully applied to assess localized degradation from invasive species and wetland loss, its application in intensive agricultural landscapes remains vital for ensuring long-term food stability [10,11]. There has also been an application of large-scale land degradation monitoring combined with drought analysis, which in turn has led to recommendations for climate-smart agriculture [12]. However, to address the land degradation status and interpretation of those three sub-indicators accurately, supporting information regarding the socio-ecological system, both socio-economic as well as cultural information, specifically for the study area, is necessary [13]. As land degradation reflects both ecological conditions and human use, assessments must integrate biophysical and socio-economic factors [9].

Rice serves as the cornerstone of global nutrition, with Asia accounting for 90% of its total production and consumption [14]. Within this regional context, Thailand plays a critical role in global food security, with annual paddy production averaging 31–32 million tonnes (equivalent to about 20–21 million tonnes of milled rice), ranking it among the world’s leading rice exporters [15,16]. In addition, the rice industry in Thailand accounts for about 5.1 million households, or 65.8% of those in the agricultural sector [15]. Therefore, at the provincial level, rice cultivation is a primary driver of the local economy and a pillar of domestic food stability. To meet high demand, Thai agricultural systems often utilize intensive monoculture, involving the repeated cultivation of rice on the same land through double- or triple-cropping cycles. To maintain productivity in these systems, farmers often resort to excessive fertilization and irrigation. Over decades of monocropping and chemical-intensive farming, the government has raised concerns about land degradation, which in turn may lead to higher product costs and lower yields for farmers. Moreover, these practices can lead to severe complications, including nitrogen pollution of groundwater and depletion of soil biodiversity [17]. Such factors diminish the land’s productive capacity and economic value, whether temporarily or permanently. Although long-term monoculture of rice cultivation is often associated with land degradation, evidence suggests that degradation is not inherent to monocropping systems but is largely driven by unsustainable management practices [18,19]. For instance, rising concerns over the intensive use of agrochemicals have led to the promotion of Good Agricultural Practices (GAP). Farmers are encouraged to combine organic and inorganic fertilizers to sustain crop quality [19,20].

This study assesses the land degradation status and sub-indicator trajectories within long-term rice monoculture systems. To overcome the challenges of time-consuming, labor-intensive, and costly field assessments, this study utilizes remote sensing-based methods within the SDG 15.3.1 indicator framework. This approach provides a scalable and efficient analysis of land degradation trajectories. In addition, qualitative field verifications and semi-structured interviews with local farmers and provincial agricultural officers are integrated to provide a more comprehensive interpretation of land degradation. The findings will be used to elaborate on the potential degradation risks in intensive rice systems and provide evidence-based recommendations for enhancing sustainable cultivation practices.

2. Materials and Methods

2.1. Study Area

Rice cultivation prevails as the predominant agricultural activity in Thailand, encompassing almost half of the total agricultural land. This study focuses on Nakhon Sawan province, distinguished for being the largest rice-producing region in northern Thailand, which has a substantial capacity for rice production, estimated at approximately 1.7 million tonnes [21]. Geographically situated at 15.6978° N, 100.1200° E, Nakhon Sawan province comprises 15 districts and spans a total area of 0.99 million hectares. The land-use composition reveals that 78% of the area is dedicated to agriculture, while forests, built-up areas, water bodies, and miscellaneous categories constitute 10%, 7%, 3%, and 2%, respectively (Figure 1). Rice plantations cover 56% of the agricultural land. The confluence of the Ping and Nan Rivers within Nakhon Sawan forms the Chao Phraya River, and the region experiences an annual precipitation ranging from 1000 to 1200 mm [22].

Rice cultivation in Thailand is divided into two distinct seasons, viz., the in-season and the off-season. In-season rice farming, also known as “major rice”, primarily occurs during the rainy season, starting in May and extending until February of the subsequent year, with harvesting concluding in April. On the other hand, off-season rice farming, or second rice farming, involves rice cultivation during the dry season, beginning in November and lasting until June of the following year, with harvesting concluding in October [23]. Nakhon Sawan province plays a vital role in both in-season and off-season rice farming, relying on precipitation during the rainy season and employing an irrigation system facilitated along the Chao Phraya River during the dry season.

2.2. Data Sources

Following the SDG 15.3.1 framework, land use and land cover (LULC), soil organic carbon (SOC), and land productivity were selected for a comprehensive assessment of land degradation in Nakhon Sawan’s rice monoculture system. Detailed data information and sources can be found in

Table 1. Data sources and land degradation parameters assessment.

Sub-Indicators	Data	Resolution	Period	Description
LULC	Land-use information (polygon) a	-	2007–2020	Provides land-use type in Nakhon Sawan.
SOC stock change	Annual soil organic carbon (OpenLandMap) b	1 km	2007–2018	Provides annual soil organic carbon (g kg−1) at 30 cm soil depth.
	Soil bulk density (OpenLandMap) c	250 m	2007–2018	Provides average soil bulk density ($10\times$ kg m−3) at 30 cm soil depth
Land productivity	Net Primary Production (NPP) d	500 m	2007–2022	The MOD17A3HGF provides information about annual NPP.

Note that a, b, c, and d refer to data sources accessible online through weblinks provided as follows: ^a https://tswc.ldd.go.th/DownloadGIS/Index_Lu.html (accessed on 15 September 2021). ^b https://stac.openlandmap.org/log.oc_iso.10694/collection.json?.language=en (accessed on 5 January 2026). ^c https://stac.openlandmap.org/bulkdens.fineearth_usda.4a1h/collection.json?.language=en (accessed on 5 January 2026). ^d https://ladsweb.modaps.eosdis.nasa.gov/ (accessed on 26 February 2026).

. The LULC data for this study were sourced from the Land Development Department [24], which provided information for the years 2007 and 2020. These data were processed in a Geographic Information System (GIS) using QGIS software version 3.38.1, where the land-cover types were classified into six categories: forest land, grassland, cropland, wetland, settlements, and other land. The statistical software (Python version 3.14) was used for Mann–Kendall trend analysis, while Microsoft Excel was used for state and performance productivity analysis. The long-term rice monoculture area was defined through a series of operations, including overlay, dataset joining, and clipping. To ensure consistency across datasets, all data were ultimately resampled to a spatial resolution of 500 m. The vector-based annual land-use datasets were converted into raster format at a 500 m spatial resolution using the Majority resampling algorithm. This grid cell assignment ensured that each pixel retained the dominant land-cover category within its spatial boundary, making it highly suitable for discrete categorical LULC analysis and minimizing edge-effect area distortions.

To calculate the soil organic carbon (SOC) stock changes, this study utilized two primary datasets: annual soil organic carbon concentration [25] and soil bulk density [26]. The data was obtained from OpenLandMap, which predicts soil properties by combining satellite imagery with ground measurements using machine learning [27]. Due to limited data availability, the study period was restricted to the years 2007 through 2018. The 1 km spatial resolution for SOC data was selected in accordance with the UNCCD Good Practice Guidance for SDG indicator 15.3.1. This resolution is sufficient for regional sub-national assessments to capture macro-level soil degradation while maintaining temporal consistency across the study period. The SOC data were in concentration units (g kg⁻¹), which required conversion into SOC stocks (t ha⁻¹). This conversion was performed by integrating information on soil bulk density, depth, and the volume of coarse fragments. For this specific analysis, the calculation focused on a soil depth of 30 cm and proceeded under the assumption that no coarse fragments were present in the soil profile.

The land productivity data were obtained from the MODIS/Terra Net Primary Production (MOD17A3HGF) dataset [28]. To align with the SDG 15.3.1 guidelines, which require a 16-year monitoring period to evaluate trends, the study period covers the years 2007 to 2022. All datasets were processed in QGIS by overlaying the productivity data with the long-term rice monoculture areas identified during the LULC assessment. Generative Artificial Intelligence (GenAI) was used for the purpose of language editing of the manuscript.

To harmonize the variations in temporal windows across multi-source datasets (LULC: 2007–2020; SOC: 2007–2018; land productivity: 2007–2022), a relative trajectory approach was applied in accordance with the UNCCD Good Practice Guidance. All sub-indicators were bound to a standardized baseline year of 2007. Slow-evolving variables like SOC (2007–2018) capture net decadal stock shifts, while fast-evolving climate-sensitive metrics like productivity (2007–2022) and ground-truth yield patterns utilize the maximum available chronological records to mitigate short-term annual anomalies. Moreover, to support the remote sensing metrics, qualitative field verifications were integrated into the study area. Purposive semi-structured interviews were conducted with select local rice farmers and provincial agricultural officers in Nakhon Sawan. These discussions focused on establishing a baseline understanding of regional agricultural practices—specifically cropping frequency, mechanical tillage intensity, agrochemical usage, and straw residue management. This ground-level context was utilized qualitatively to interpret and ground-truth the spatial degradation trends and localized hotspots identified by the satellite data.

2.3. Qualitative Field Observation via Semi-Structured Interview

Qualitative field validation was conducted using semi-structured interviews to bridge the gap between large-scale satellite data and actual ground realities of rice plantation in Nakhon Sawan province. Regional agricultural officers selected five local rice farmers representing five districts of this province, i.e., Phai Sali, Krok Phra, Phayuha Khiri, Banphot Phisai, and Lat Yao, by using a purposive sampling approach. Rice cultivation in these areas is a deeply rooted generational practice passed down through families, characterized by highly stable, long-term operations with minimal transition to other crops due to persistent seasonal flooding and specialized irrigation infrastructure.

The interviews revealed that local farmers primarily rely on transplanting and broadcast-seeding methods, often incorporating the Alternate Wetting and Drying (AWD) water management technique. The standard seasonal timeline for major rice cultivation typically begins between May and July, with a growth cycle lasting around 120 days. In the paddy fields in the irrigation or groundwater reservoir areas, farmers can increase single-cropping cycles to double- or triple-cropping cycles per annum due to the water availability. Chemical fertilizers are mainly used in the proportion of nitrogen (N), phosphorus (P), and potassium (K) at 16-20-0, 16-8-8, and 46-0-0 (urea); these are applied at critical growth stages, specifically around days 20, 45, 55, and 75. In addition, tailor-made fertilizers are also utilized in some paddy fields, which are designed based on soil quality and nutrients analysis from Regional agricultural officers. The rice biomass, e.g., rice stubble, is left in the field, which is then turned back into the soil through the plowing and tillage process.

2.4. Land Degradation Assessment

2.4.1. Land-Use and Land-Cover (LULC) Degradation Assessment

To evaluate changes in LULC, this study compared land-use status between the baseline year (2007) and the target year (2020) using Nakhon Sawan’s land-use data with a spatial resolution of 500 m. Subsequently, the land-cover change was calculated between two years and generated a land-cover transition matrix. Finally, the land-cover change trends were determined by categorizing them as improved, degraded, or stable based on Figure 2. Due to the limited availability of comprehensive annual land-use datasets from public institutions, consecutive annual maps for the entire intervening period (2008–2019) could not be fully integrated into spatial modeling. To overcome this limitation and validate the temporal persistence of the identified paddy areas, field-level verification was conducted through interviews with local farmers and provincial agriculture officers. The ground-level evidence confirmed that rice cultivation within these core zones is a multi-generational practice passed down through families, characterized by highly stable long-term operations with minimal transition to other crops due to seasonal flooding and specialized irrigation infrastructure. Thus, these zones are designated as persistent paddy rice areas across the study timeline.

2.4.2. Soil Organic Carbon (SOC) Stock Degradation Assessment

The SOC stock value was simulated for both 2007 and 2018, and the relative change was calculated. Areas exhibiting a decline in SOC stock greater than 10% were classified as degraded, whereas those showing an increase greater than 10% were considered improved. The 10% SOC change threshold implemented in this analysis strictly adheres to the international standardized baseline mandated by the UNCCD Good Practice Guidance for SDG Indicator 15.3.1.

2.4.3. Land Productivity Degradation Assessment

The assessment of land productivity is conducted based on three key metrics, i.e., trend, state, and performance. The productivity trend describes the trajectory of productivity variations throughout the study duration. It is computed by utilizing annual productivity measurements spanning from 2007 to 2022. The productivity state is determined by comparing the average annual NPP of the three most recent years with the range of annual NPP values observed in the preceding 13 years.

The trend of productivity is determined by assessing the changing trend of NPP over time. The Mann–Kendall non-parametric significance test is applied on the time series data of NPP, pixel by pixel [29,30,31]. Equation (1) is used for the trend of productivity calculation.

$$S={\displaystyle {\sum }_{i=1}^{n-1}}{\displaystyle {\sum }_{j=i+1}^n}sgn(x_j-x_i)$$

(1)

where sgn(x_j − x_i) values fall between −1 and 1, as presented in Equation (2).

$$sgn\left(x_j-x_i\right)=\left\{\begin{array}{c}1 if(x_j-x_i>0)\\ 0 if(x_j-x_i=0)\\ -1 if(x_j-x_i<0)\end{array}\right.$$

(2)

Here, x_i and x_j are the time series values (annual NPP), and n is the number of years.

Under the null hypothesis of no trend, the test statistic S is approximately normally distributed for n > 10, with the mean and variance given by Equation (3) and Equation (4), respectively.

$$E\left(S\right)=0$$

(3)

$$Var\left(s\right)={\displaystyle \frac{n(n-1)(2n+5)}{18}}$$

(4)

The standardized test statistic Z is estimated through Equation (5).

$$Z=\left\{\begin{array}{c}{\displaystyle \frac{S-1}{\sqrt{var\left(s\right)}}} S>0\\ 0 S=0\\ {\displaystyle \frac{S+1}{\sqrt{var\left(s\right)}}} S<0 \end{array}\right.$$

(5)

The productivity state assesses changes in NPP between the baseline period (2007–2019) and the comparison period (2020–2022). Significance is determined by calculating the Z value using Equations (6) and (7) [8,32].

$$\sigma =\sqrt{{\displaystyle \frac{{\sum }_{i=1}^N{(X_i-\overline{X})}^2}{N}}}$$

(6)

$$z={\displaystyle \frac{\overline{X}-\mu }{\raisebox{1ex}{$\sigma $}\!\left/ \!\raisebox{-1ex}{$\sqrt{3}$}\right.}}$$

(7)

wherein σ is the standard deviation; N is the number of years in the baseline year; $\overline{X}$ represents the mean of annual NPP of the observed years; μ denotes the mean of the baseline year. The analysis of trend and state uses a Z-score threshold of less than −1.96 to identify degraded areas (p-value < 0.05), while a Z-score exceeding 1.96 suggests significant improvement (Figure 3).

Productivity performance is evaluated by calculating the mean NPP value for each pixel and employing the 90th percentile NPP value as the maximum productivity threshold for each unit. Subsequently, the ratio between the average NPP of each pixel and the maximum productivity of the ecological unit is computed, with pixels falling below 0.5 designated as degraded [8].

In the integrated interpretation, primary significance is assigned to the performance indicator; if a pixel exhibits degradation in performance, it is classified as degraded. Conversely, if performance is not degraded, the assessment then considers both trend and state, where a degraded status is only assigned if both indicators concurrently exhibit decline (Figure 3).

Following the SDG 15.3.1 framework, three independent sub-indicators—LULC, SOC stock, and land productivity—were integrated to assess overall degradation using the “One-Out, All-Out” principle. The overarching “One-Out, All-Out” principle is applied solely at this integration stage, dictating that overall land degradation is triggered if any of the three main indicators registers degradation.

3. Results and Discussion

3.1. Land-Use and Land-Cover (LULC) Change Degradation Assessment

Land cover in Nakhon Sawan is categorized according to the IPCC guidelines, which distinguish six primary classes: forest land, grassland, cropland, wetland, settlements, and other land. However, given the study’s focus on long-term rice monoculture, the cropland category was further subdivided into paddy rice and other croplands for a more detailed classification of agricultural practices. The land-use and land-cover change in Nakhon Sawan’s rice monoculture was analyzed from 2007 to 2020. As of 2007, paddy rice cultivation accounted for 4011.67 km², representing 42% of the total land area. Figure 4 illustrates the comparative land cover and subsequent identification of land use. To assess land degradation specifically within long-term intensive agriculture, this study focused only on the area of long-term rice monoculture. By focusing on pixels that remained as rice paddies throughout the study period, the influence of land-use conversion was eliminated, allowing for a direct assessment of degradation within the rice cropping system itself.

The analysis considered a 500 m grid resolution to isolate these persistent rice-growing areas from the other land-use types. A spatial overlay was performed between the baseline year (2007) and the target year (2020) to identify the intersection of rice-growing areas. This target area serves as the primary zone for evaluating land productivity and soil organic carbon dynamics within the rice monoculture system. The total area dedicated to paddy rice cultivation contracted progressively, reflecting a strategic shift toward diversified crop production. The paddy rice cultivation transitioned to sugarcane, followed by cassava, corn, other field crops, perennials, and orchards. The remaining area transitioned into settlements, wetland, grassland, as well as miscellaneous land-use types such as abandoned land, pits, landfills, and mining sites. By 2020, rice monoculture remained the dominant practice, comprising 83.12% of all paddy rice cultivation (

Table 2. Land-use transition matrix for paddy rice in Nakhon Sawan (2007–2020).

LULC 2007	LULC 2020 (Transition Categories)
	Paddy Rice	Other Croplands	Grassland	Settlements	Wetland	Other Lands
Paddy rice	3304.25 km2	509.75 km2	6.25 km2	53 km2	30.25 km2	71.75 km2
	83.12%	12.82%	0.16%	1.33%	0.76%	1.80%

Remark: Yellow = stable, light red = degradation.

).

To quantify the land degradation, a land-cover transition matrix was employed to identify the directional shifts between categories. This matrix highlights the conversion of productive cropland into degraded states, such as grassland, wetlands, settlements, and other land. The land status is determined by land-use and land-cover change transfer rules.

Based on the observed transitions in paddy rice cultivation, land degradation resulting from land-cover change was estimated at 16,125 ha, representing 4.05% of the paddy rice cultivation area. Following the established transition rules, the area under paddy rice monoculture revealed a stable land status with no detectable transitions during the study period. No paddy rice cultivation area shifted to forest land during the study period; consequently, no land improvement was recorded in this category.

3.2. Soil Organic Carbon (SOC) Stock Changes: Degradation Assessment

Soil organic carbon (SOC) is a critical indicator of soil health and carbon sequestration capacity. Within the framework of SDG 15.3.1, a significant decrease in SOC stock is interpreted as a sign of land degradation. This study details the SOC dynamics specifically within the area of long-term rice monoculture from 2007 to 2018. Following the IPCC Tier 1 methodology, the SOC stocks were calculated based on the 500 m grid resolution. The assessment focused on the top 30 cm of the soil profile, where agricultural management practices most directly influence carbon levels. The percentage of SOC stocks changes within the long-term rice monoculture is presented in Figure 5. Areas marked red represent the rice monoculture systems in which SOC stocks exhibited a decline to 14.3%, the yellow represents no change (0%), whereas green areas showed potential carbon accretion trends of up to 35.3%.

The analysis of SOC stock changes across different districts in Nakhon Sawan provides a granular view of land degradation hotspots. While some districts showed resilience, others experienced carbon depletion within their long-term rice monoculture zones. Specifically, districts such as Phayuha Khiri, Takhli, Mueang Nakhon Sawan, and Krok Phra experienced significant SOC decrease, signaling a decline in soil quality. Conversely, a predominant potential carbon accretion trend was observed in Nong Bua, Banphot Phisai, and Phai Sali, suggesting more effective organic matter retention in these areas (

Table 3. Summary of soil organic carbon (SOC) stock changes by district.

District	The Proportional Area of SOC Stock Changes
	Decrease ($∆\mathit{S}\mathit{O}\mathit{C}<0\mathit{\%})$	Stable ($∆\mathit{S}\mathit{O}\mathit{C}=0\mathit{\%})$	Increase ($∆\mathit{S}\mathit{O}\mathit{C}>0\mathit{\%})$
Banphot Phisai	1.4%	7.9%	90.7%
Nongbua	0.2%	4.8%	95%
Chum Saeng	0%	3.8%	96.2%
Kao Liew	0.6%	2.8%	96.7%
Mae Wong	3.8%	20.5%	75.7%
Lat Yao	0.7%	23.3%	76%
Mueang Nakhon Sawan	9.3%	37.3%	53.4%
Tha Ta Ko	3.4%	25.2%	71.4%
Phai Sali	1.3%	16%	82.7%
Chum Ta Bong	2.4%	3.9%	93.7%
Krok Phra	21%	34.5%	44.5%
Phayuha Khiri	48.1%	37.1%	14.9%
Tak Fa	0%	33.3%	66.7%
Ta Khli	22.2%	31.5%	46.3%

).

The data revealed a clear spatial divide in soil health across the province. The high decrease in carbon in districts like Phayuha Khiri suggests that these areas may be under higher intensive farming pressure or have soil characteristics that are more susceptible to carbon mineralization. In contrast, almost every district showed a higher proportion of areas with an increase in SOC stocks than a decrease in SOC stocks. Given the global scale of this regional screening, high-resolution quantitative datasets tracking localized fertilizer application or irrigation volumes were unavailable at the farm level. Consequently, the apparent spatial associations between the positive SOC trajectories and agricultural zones should be interpreted as diagnostic hypotheses rather than explicit empirical causations. To logically address this data gap, these spatial trends were thoroughly cross-verified with rice research experts. These SOC increase results are likely attributable to post-harvest management practices; typically, rice stubble is left in the paddy fields to decompose, serving as an important source of organic matter that maintains and enhances SOC stocks. This is further supported by government initiatives encouraging the substitution of chemical fertilizers with organic alternatives. Liu et al. [33] indicate that returning biomass to the soil increases carbon stocks by 13%. Furthermore, the application of organic fertilizers, either alone or combined with NPK, has been shown to increase SOC stocks by 19% and 32%, respectively. Sharma et al. [34] observed that long-term rice monocropping facilitates the accumulation of high organic carbon content due to a low rate of decomposition. Specifically, the long-term practice of partial residue burning followed by soil puddling, which enhances organic carbon stocks in both surface and subsurface horizons. The resulting partially burned residue acts as a form of natural biochar, stabilizing soil health and mitigating further degradation even within intensive puddled systems. Therefore, the gains in many districts such as Banphot Phisai, Chum Saeng, Nongbua, and others, provide a baseline for successful soil management that could be studied and replicated in degraded zones.

In accordance with SDG 15.3.1 guidelines, land is classified as “degraded” with respect to soil organic carbon (SOC) only when stocks exhibit a reduction exceeding the 10% significance threshold relative to the baseline. Even though a province may look stable overall at a large (provincial) scale, this broad picture can hide severe environmental damage and declining soil health in specific areas (at the district level), as shown in Figure 6. Applying this criterion to long-term rice monocultures in Nakhon Sawan reveals that Phayuha Khiri, Takhli, Krok Phra, and Mueang Nakhon Sawan contain the most extensive degraded areas, covering 4, 1.5, 0.5, and 0.25 km², respectively. It is worth noting that while 48.1% of the persistent paddy area in Phayuha Khiri district exhibited a negative soil carbon trajectory (∆SOC $<$ 0%), the localized area that experienced severe SOC depletion crossing the official SDG 15.3.1 degradation threshold ($∆\mathrm{SOC}\le$ −10%), was confined to 4 km². This demonstrates that while a minor decrease in carbon is widespread across the district’s intensive cropping systems, critical land degradation is concentrated within specific high-stress agricultural pockets. While many areas experienced SOC decreases between 0% and 10%, these remain technically categorized as “stable”. This distinction is essential for prioritizing policy interventions, ensuring that restoration of resources is targeted toward districts where degradation surpasses the reporting threshold.

3.3. Land Productivity Sub-Indicator Degradation Assessment

To assess the degradation status of land productivity in the long-term rice monoculture from 2007 to 2022, this study analyses the changes in the three sub-indicators of land productivity: trend, state, and performance. By using the 500 m grid resolution, the trend and state categorized the productivity into five standard classes, i.e., degrading, potentially degrading, stable, potentially improving, and improving. While performance was determined by comparing the productivity of a pixel to the maximum NPP, most of the long-term rice monocultures remain ‘not degraded’ for performance productivity assessment. The analysis of the long-term rice monoculture areas (Figure 7) provides a comprehensive view of productivity health.

The analysis of land productivity dynamics in Nakhon Sawan evaluates two critical components: the productivity trend, which monitors the trajectory of change over time, and the productivity state, which compares current levels against a historical baseline. As detailed in

Table 4. Trend productivity status classified by district.

District	The Proportional Area of Trend Productivity Status
	Degrading	Potentially Degrading	Stable	Potentially Improving	Improving
Banphot Phisai	0.4%	5.3%	91.2%	2.8%	0.3%
Nongbua	8.2%	29%	62.7%	0.1%	0%
Chum Saeng	0%	9.4%	87.1%	2.4%	1.1%
Kao Liew	0%	1.2%	91.1%	7.5%	0.2%
Mae Wong	2.8%	14.8%	82.1%	0.3%	0%
Lat Yao	0.1%	4.1%	92.1%	3.6%	0%
Mueang Nakhon Sawan	0%	1.8%	93.6%	4.6%	0%
Tha Ta Ko	4.1%	15.8%	79%	1.1%	0%
Phai Sali	2.9%	14.1%	81.9%	1.1%	0.1%
Chum Ta Bong	0.8%	9.9%	89.3%	0%	0%
Krok Phra	0.5%	0.7%	90.6%	7.2%	0.9%
Phayuha Khiri	3.8%	15.2%	78.2%	2.4%	0.5%
Tak Fa	0%	14.3%	85.7%	0%	0%
Ta Khli	1.5%	6.4%	82.2%	8.2%	1.7%

, several districts exhibit significant biological stress. Nong Bua presents the most concerning trajectory, with 36 km² classified as ‘degrading’ and 127.25 km² identified as ‘potentially degrading.’ Conversely, Banphot Phisai, Phai Sali, and Tha Tako demonstrate robust temporal resilience, each maintaining over 300 km² of stable land.

Analyzing the productivity state gives a clearer picture of land health in Nakhon Sawan. It shows that many areas are producing much less than they did in the past. While the trend suggests that the decline is slowing down, the state shows that the actual damage is more widespread. Specifically, Nong Bua has the largest degraded area at 102.5 km², followed by Phai Sali (66 km²) and Tha Tako (61.5 km²) (

Table 5. State productivity status classified by district.

District	The Proportional Area of State Productivity Status
	Degrading	Potentially Degrading	Stable	Potentially Improving	Improving
Banphot Phisai	9.6%	16%	74.3%	0.1%	0%
Nongbua	23.9%	24.1%	52%	0%	0%
Chum Saeng	18%	23.4%	58.6%	0%	0%
Kao Liew	2.3%	6.4%	91.3%	0%	0%
Mae Wong	23.3%	34.4%	42.3%	0%	0%
Lat Yao	13.2%	23.8%	62.7%	0.3%	0%
Mueang Nakhon Sawan	2.7%	12.1%	84.9%	0.3%	0%
Tha Ta Ko	14.6%	26.8%	58.5%	0.1%	0%
Phai Sali	15.8%	21.3%	62.8%	0.1%	0%
Chum Ta Bong	25.2%	24.4%	50.4%	0%	0%
Krok Phra	4.3%	15%	80.5%	0.2%	0%
Phayuha Khiri	11.8%	24.4%	63.8%	0%	0%
Tak Fa	24.2%	33.3%	42.4%	0%	0%
Ta Khli	5.1%	12.7%	81.3%	0.6%	0.3%

). In almost every district, the amount of “degrading” land is much higher when looking at the current state rather than the trend. This means that even though the soil is getting worse more slowly, its overall health is still much lower than it used to be.

For a better understanding of the dynamics of land productivity, we analyzed the factors that influence NPP. The analysis of rice NPP reveals that climatic parameters, including solar radiation, temperature, soil moisture, and the fraction of photosynthetically active radiation (FPAR), are essential requirements for productive rice cultivation. Solar radiation acts as the primary energy source for NPP dynamics, showing a consistently positive relationship where higher radiation levels correspond to increased productivity. Huailuek et al. [35] studied changes in paddy rice NPP using the Carnegie–Ames–Stanford Approach (CASA) Model for the years 2007–2018, revealing that the annual mean NPP was 832, 808, 825 gC m⁻² year⁻¹, respectively (

Table 6. Characteristics of climatic parameters.

Months	NDVI	FPAR	Water Factors	Solar Radiation (MJ)	Sunshine Duration (Hours)	Temperature (°C)	Effective Rainfall (mm)	Precipitation (Days)	NPP (gC m−2)
Average Values (unit/month)
2007	0.52	0.49	0.98	553	217	28.4	64	11	69.3
2012	0.53	0.50	0.99	535	198	29.1	62	11	67.4
2018	0.54	0.49	0.99	545	205	28.2	49	9	68.8
Monthly Values
January	0.43–0.50	0.46–0.50	0.98–0.99	425–475	193–262	25.3–26.7	0–18	2–5	51.1–54.0
February	0.40–0.48	0.47–0.50	0.91–1.00	458–520	221–275	26.4–28.7	0–26	1–4	56.0–58.6
March	0.36–0.45	0.46–0.48	0.96–0.99	574–600	174–267	29.8–30.6	0–15	1–5	66.0–73.4
April	0.40–0.46	0.47–0.48	0.97–0.98	537–651	221–268	29.9–31.7	40–78	5–11	64.8–79.4
May	0.46–0.56	0.48–0.51	0.99–1.00	565–667	180–202	29.4–30.1	98–141	16–22	74.3–84.6
June	0.55–0.58	0.50–0.51	0.99–1.00	541–605	143–200	28.1–29.4	72–99	13–23	70.6–88.4
July	0.48–0.57	0.50–0.51	0.99–1.00	502–535	118–184	28.1–28.7	83–132	16–20	65.1–78.3
August	0.52–0.60	0.49–0.51	0.99–1.00	488–549	138–169	28.1–28.6	51–149	17–22	62.8–70.5
September	0.56–0.63	0.49–0.54	0.99–1.00	470–546	141–203	27.7–28.5	44–159	11–20	58.5–75.0
October	0.63–0.68	0.53–0.54	0.98–1.00	507–573	185–203	28.1–29.1	17–129	5–17	71.0–79.0
November	0.56–0.63	0.48–0.51	0.97–1.00	452–496	195–267	27.3–29.1	0–45	3–6	58.3–60.1
December	0.47–0.56	0.47–0.49	0.96–0.99	507–617	226–287	26.9–29.1	0–13	0–6	59.0–76.7

). These values are higher than NPP from field measurements in Naser et al. [36], which ranged from 499 to 530 gC m⁻² year⁻¹. The difference occurs because the field measurements were limited to the cultivation period, whereas the model integrates monthly parameters to determine the total annual value. However, the purpose of this calculation is to examine the relationship between the climatic parameters and the NPP value. The results show that 2007 had the highest annual mean NPP. This is because solar radiation was stronger in 2007 than in the other two years. On a monthly scale, May was identified as receiving the highest solar radiation, contributing to it being the most suitable month for rice cultivation.

Temperature plays a critical role in determining light use efficiency, with the study identifying an optimal growth temperature of 28.6 °C. The relationship between temperature and NPP follows a quadratic correlation, where both excessively high and low temperatures lead to a decline in productivity. Elevated temperatures are particularly risky as they can increase respiration rates to the point where they exceed the rate of photosynthesis, resulting in a rapid consumption of photosynthates and an overall reduction in NPP. Consequently, months like January, February, and March often exhibit lower NPP values because their temperatures differ significantly from the optimal range.

Water availability and vegetation greenness further modulate these productivity levels. Precipitation shows a robust association with NPP, as intensified rainfall increases soil moisture and amplifies photosynthesis. While the study area maintained high soil water factors (0.9 to 1) due to the substantial water demands of rice, the highest NPP values were specifically noted during wet months like May, June, and October. Beyond the volume of rainfall, the number of precipitation days was found to be a significant factor in maintaining productivity. These climatic factors work in tandem with the rice’s growth phases: seedling, vegetation, generative, and fallow, which are monitored via NDVI and FPAR to ensure the plants are efficiently utilizing light and energy throughout their lifecycle.

The analysis of rice monocultures in Nakhon Sawan reveals a clear link between soil health and climate patterns. While most of the study areas are stable, districts like Tha Tako and Phayuha Khiri have emerged as degradation hotspots. These areas suffer from a “double stress” phenomenon: they have lost more than 10% of their SOC and are also seeing a decline in productivity. This drop is largely caused by heat stress. Rice grows best at an optimal temperature of 28.6 °C, but as temperatures rise above 30 °C, the plants begin to consume energy faster than they can produce through photosynthesis. This imbalance leads to a rapid drop in NPP, explaining why even well-managed farms show signs of degradation during extreme heat.

We found a different situation in the Nong Bua district. The soil carbon has increased, yet productivity continues to fall. This highlights that soil nutrients alone are not enough; sunlight and water are also critical limiting factors. The high productivity levels recorded in 2007 were driven by peak solar radiation, a level that was not reached in 2012 or 2018. In Nong Bua, the gains in soil carbon, likely from historical flood sediments, could not offset the impact of severe drought. When rainfall drops below 10 mm during the early stages, rice plants cannot use light efficiently, regardless of how healthy the soil is.

3.4. Rice Monoculture Land Degradation Assessment and Spatiotemporal Analysis of Rice Yield

The integrated assessment of land degradation in Nakhon Sawan’s persistent rice-growing areas, conducted under the SDG 15.3.1 framework, combined three sub-indicators: LULC, SOC stock, and land productivity. The first sub-indicator, LULC, indicated a “not degraded” status because the study focused on long-term rice monocultures, avoiding conversion to less productive land types. The second sub-indicator, SOC stock changes, showed that most of the area is “not degraded”, with the majority of the region experiencing an overall increase in SOC stocks.

The third sub-indicator, land productivity, integrated three metrics: trend, state, and performance. While the performance metric is classified as “not degraded”, both the trend and state metrics revealed that biological productivity is partially or potentially degraded. When these three metrics are combined, only a small fraction of the total area is identified as degraded (Figure 7d). Following the “One-Out, All-Out” principle of the SDG 15.3.1 framework, almost all of Nakhon Sawan’s long-term rice monoculture areas are classified as “not degraded”. Only small portions of the total study area were identified as degraded (Figure 8).

To verify the results from the SDG 15.3.1 framework, which are indicated as “non-degraded” land, a linear regression analysis was performed on yield data for all 15 districts in Nakhon Sawan over a 15-year period (2009–2023) to evaluate long-term productivity shifts, as shown in

Table 7. Fifteen-year Nakhon Sawan’s rice yield analysis by district.

District	System	Rice Yield (kg ha−1)				Slope (Trend)	p-Value
		Mean	Max	Min	CV (%)
Banphot Phisai	Major	3319	3994	2500	14.18	−18.79	0.5385
	Second	4513	5231	3750	9.84	−26.07	0.3615
Nongbua	Major	3195	3913	2581	13	−46	0.0712
	Second	4436	5163	3831	9.33	−42.39	0.0988
Chum Saeng	Major	3334	3906	2581	13.23	−45.27	0.0979
	Second	4632	5269	3831	9.72	−52.23	0.057
Kao Liew	Major	3142	3863	2525	12.52	−6	0.8153
	Second	4461	5113	3775	9.19	+3.42	0.8986
Mae Wong	Major	3083	3963	2575	14.42	−33.15	0.2417
	Second	4428	5213	3756	10.39	−2.12	0.9438
Lat Yao	Major	3148	4025	2556	14.54	+32.88	0.2602
	Second	4442	5275	3825	10.48	+43.55	0.1353
Mueang Nakhon Sawan	Major	3169	4056	2500	17.58	−8.35	0.8187
	Second	4387	5250	3750	12.14	−14.6	0.6742
Tha Ta Ko	Major	3274	3925	2588	15.83	−9.4	0.7814
	Second	4559	5194	3838	11.39	−33.33	0.3172
Mae Poen	Major	3224	3725	2594	10.05	+13.42	0.5235
	Second	4586	5163	3888	8.38	−34.31	0.1559
Phai Sali	Major	3289	3950	2588	12.87	−34.78	0.1942
	Second	4485	5200	3813	9.23	−26.32	0.3218
Chum Ta Bong	Major	3372	3925	2569	13.15	+2.68	0.9264
	Second	4675	5200	3819	9.62	+0.80	0.9782
Krok Phra	Major	3225	4006	2569	15.63	−50.27	0.1089
	Second	4333	5225	3819	9.87	−28.86	0.2918
Phayuha Khiri	Major	3347	3888	2550	13.74	−3.35	0.9114
	Second	4491	5138	3800	9.11	+26.25	0.3173
Tak Fa	Major	3243	4044	2506	15.65	−47.9	0.1315
	Second	4548	5294	3756	11.41	−50.18	0.121
Ta Khli	Major	3274	3850	2594	11.14	−4.84	0.839
	Second	4551	5100	3925	7.3	−11.18	0.6049

. Longitudinal analysis shows clear differences between major (in-season) and second (off-season) rice systems. Major Rice had a lower mean yield (3243 kg ha⁻¹) compared to Second Rice (4500 kg ha⁻¹) and higher variability (CV = 13.8% vs. 9.8%), indicating lower stability. This is due to its reliance on natural rainfall and vulnerability to monsoon variability and flooding. In contrast, Second Rice is more resilient due to controlled irrigation from the Chao Phraya river basin, which reduces climate-related risks.

For long-term productivity shifts, the results revealed that there was no statistically significant downward trend in rice yields across any district in Nakhon Sawan when subjected to a 5% significance test (p > 0.05). While not reaching the threshold of statistical significance, it was found that the districts such as Nongbua and Chum Saeng had marginal negative slopes with p-values ranging between 0.05 and 0.10. These areas represent critical “points of interest” for policymakers, especially for major rice cultivation in Nong Bua, as the downward pressure on yields—though currently non-significant—may escalate under future climate scenarios, compared to the specific indication of land degraded as “red area” of Figure 8, which is mainly located in Nongbua district, followed by Tha Ta Ko and Phai Sali.

The lack of significant degradation despite long-term monoculture of rice cultivation in Nakhon Sawan could be the result of several factors. Paddy rice cultivation involves flooding, which protects soil carbon from rapid oxidation [37]. Furthermore, sustainable management practices, such as stubble retention and the incorporation of organic amendments, can mitigate degradation effects for decades [33,34]. Nakhon Sawan is located in the Chao Phraya River basin, which allows for annual alluvial deposits. This siltation acts as a natural fertilizer, replenishing minerals that are harvested with grain. In addition, the development and use of high-yield variety also compensate for land degradation.

In conclusion, the SDG 15.3.1 framework is a well-established indicator for land degradation assessment [3,4,38] and is highly appropriate for evaluating agricultural landscapes [11]. This approach is significantly less time-consuming and labor-intensive than traditional field examinations; however, integrating information from field observations can further strengthen the reliability of the results.

4. Conclusions

The study of Nakhon Sawan’s long-term rice monoculture indicates that while land-use and land-cover (LULC) transitions led to degradation in about 4% of the paddy rice area, the core areas of monoculture remained largely stable. The analysis of soil organic carbon (SOC) stocks revealed a complex picture of soil health; despite significant carbon depletion in districts like Phayuha Khiri and Takhli, a broader potential carbon accretion trend was observed across nearly every district. This overall increase in SOC stocks is likely linked to post-harvest management practices, such as retaining rice stubble and utilizing organic amendments, which help maintain soil quality and mitigate long-term degradation. While land productivity metrics like “trend” and “state” highlighted some productivity stress, particularly in Nong Bua, the overall performance remained stable. Ultimately, the application of the SDG 15.3.1 framework demonstrated that long-term rice cultivation in this region does not result in widespread land degradation, provided sustainable management practices are integrated into the agricultural system. This study pointed out the vulnerable hotspot areas with potentially degraded productivity and SOC depletion, which serve as a warning signal for agricultural production in this region. However, the identified district-level degradation hotspots should be explicitly contextualized within the Agricultural Climate Change Strategic Plan to transition these technical spatial metrics into policy-relevant tools. For the high-stress zones in Phayuha Khiri and Takhli districts, where satellite data revealed severe declining SOC trajectories (∆SOC $\le$−10%), the findings mandate a policy shift toward “Zero-Burning Circular Agriculture” through the incorporation of rice stubble as a soil amendment. For future research, the study should examine the soil data combined with agricultural management to clarify the specific local “Degraded” area. Although the global default data can evaluate the status of land degradation, local soil measurement is still required for comprehensive land assessment. The agricultural management of both “Degraded” and “Stable” areas should be studied for successful soil management and the prevention of land degradation. Furthermore, a spatial constraint of this study is its reliance on a 500 m resolution across all analytic layers. This resolution was selected to match the MODIS Net Primary Productivity (NPP) dataset (MOD17), which was utilized for analyzing land productivity trajectories. This macro-scale screening resolution leads to an inherent spatial smoothing effect, which inevitably averages out finer sub-pixel heterogeneities. While this 500 m resolution remains effective for regional-level diagnostics and policy target-screening, integrating higher-resolution data presents an important pathway for future field-scale validations.