*Article* **Assessing Soil Organic Carbon, Soil Nutrients and Soil Erodibility under Terraced Paddy Fields and Upland Rice in Northern Thailand**

**Noppol Arunrat 1,\*, Sukanya Sereenonchai <sup>1</sup> , Praeploy Kongsurakan <sup>2</sup> and Ryusuke Hatano <sup>3</sup>**


**Abstract:** Terracing is the oldest technique for water and soil conservation on natural hilly slopes. In Northern Thailand, terraced paddy fields were constructed long ago, but scientific questions remain on how terraced paddy fields and upland rice (non-terraced) differ for soil organic carbon (SOC) stocks, soil nutrients and soil erodibility. Therefore, this study aims to evaluate and compare SOC stocks, soil nutrients and soil erodibility between terraced paddy fields and upland rice at Ban Pa Bong Piang, Chiang Mai Province, Thailand. Topsoil (0–10 cm) was collected from terraced paddies and upland rice fields after harvest. Results showed that SOC stocks were 21.84 and 21.61 Mg·C·ha−<sup>1</sup> in terraced paddy and upland rice fields, respectively. There was no significant difference in soil erodibility between terraced paddies (range 0.2261–0.2893 t·h·MJ−<sup>1</sup> ·mm−<sup>1</sup> ) and upland rice (range 0.2238–0.2681 t·h·MJ−<sup>1</sup> ·mm−<sup>1</sup> ). Most soil nutrients (NH<sup>4</sup> -N, NO<sup>3</sup> -N, available K, available Ca and available Mg) in the terraced paddy field were lower than those in the upland rice field. It was hypothesized that the continuous water flows from plot-to-plot until lowermost plot caused dissolved nutrients to be washed and removed from the flat surface, leading to a short period for accumulating nutrients into the soil. An increase in soil erodibility was associated with decreasing SOC stock at lower toposequence points. This study suggested that increasing SOC stock is the best strategy to minimize soil erodibility of both cropping systems, while proper water management is crucial for maintaining soil nutrients in the terraced paddy field.

**Keywords:** soil organic carbon; soil erodibility; terraced paddy field; upland rice; Thailand

## **1. Introduction**

Globally, the soil organic carbon (SOC) pool has been a key challenging subject since it could generate either positive or negative feedback on atmospheric CO<sup>2</sup> changes [1–3]. As soil stores three-times more carbon than the atmosphere, many land uses have endangered soil degradation and also the related SOC sequestration potential [4,5], especially under agricultural systems [6]. In agricultural land, the SOC in soils is often depleted but also has the potential to sequester carbon (C) under agricultural practices [7,8]. In addition, SOC plays a role indicator of soil quality, as it contributes to soil biochemical and physical functions that are essential for plants and microorganisms [9]. Hence, agricultural management has become increasingly highlighted as a high-potential tool for mitigating climate changes and adapting to changing climate [10,11]. Concerning agriculture in mountainous areas, the slope gradient is the key factor affecting SOC dynamics and soil quality [12–14], while soil losses and soil erosion are highly correlated with slope steepness [15]. Budry and Curtis [16] and Tadele et al. [17] indicated that soil erosion causes land degradation that

**Citation:** Arunrat, N.; Sereenonchai, S.; Kongsurakan, P.; Hatano, R. Assessing Soil Organic Carbon, Soil Nutrients and Soil Erodibility under Terraced Paddy Fields and Upland Rice in Northern Thailand. *Agronomy* **2022**, *12*, 537. https://doi.org/ 10.3390/agronomy12020537

Academic Editor: Antonio Miguel Martínez-Graña

Received: 24 November 2021 Accepted: 16 February 2022 Published: 21 February 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

is the most serious problem for the farmer in tropical highlands, resulting in decreasing soil fertility and crop yield losses. Therefore, it is challenging to investigate the variation of SOC, soil nutrients and soil erodibility under different land management techniques in hilly and mountainous areas.

The most effective engineering measure for slope management is terracing, which, when well designed and appropriately maintained, has been reported to reduce soil and water loss [18–21]. Terracing is the oldest technique for water and soil conservation on natural hilly slopes [22,23], and some terraces were constructed in Southeast Asia more than 5000 years ago [24–27]. Under the terracing system, mountainous terrains are managed into narrow graduated steps to enable crop growing and management practices [28]. The terracing system plays a vital role in soil and water conservation not only by directly creating many microtopographical and specific hydrological pathways [29–31] but also by reducing slope gradient and connectivity of overland runoff [32,33]. In terraces, divided sections of cropland with gradient slopes also help farmers to alleviate flooding and erosion, while improving other ecosystem services, including food security [34], recreation [35] and C sequestration [36]. Previous studies, however, have found varying levels of SOC in different terraced sites. Xu et al. [37] revealed that mean SOC densities at the 0–60 cm soil depth in terraces were the highest when compared to forestland, grassland and sloping cropland. Meanwhile, Zhang et al. [38] presented the idea that soil erosion and cropping contributed to variations of SOC and TN losses along the sloping terrace. Inappropriate terraces, often randomly designed by local farmers, were also marked as having a higher risk of soil erosion and severe landslides caused by the unstable structure [39,40].

In Thailand, terraced paddy fields have been widely distributed in Northern Thailand under contribution by the Royal Development Project since 2003 [41]. Even though terraced paddy fields have been determined to be a highland sustainable agriculture approach [42], the surface runoff and soil and nutrient losses need to be examined [43]. However, the investigation of SOC stock, soil nutrients and soil erodibility under rice terraces is still lacking, indicating that it would be desirable to find out how terraced paddy field and upland rice differ in SOC stock, soil nutrients and soil erodibility. To fill these gaps, our study aims to evaluate and compare (1) SOC stock and soil nutrients and (2) soil erodibility between terraced paddy fields and upland rice as well as the different potentials of rice terraces among toposequences. This study could contribute to a clearer understanding of the relevant issues of hilly and mountainous areas.

#### **2. Material and Methods**

#### *2.1. Study Area and Site Selection*

The study area was carried out at Ban Pa Bong Piang, Chang Khoeng Subdistrict, Mae Chaem District, Chiang Mai Province, Northern Thailand (Figure 1a). The rainy season of Chiang Mai Province is from May to October. The average annual temperature ranged between 20 and 35 ◦C, while the average annual rainfall ranged between 600 and >1000 mm. The topography of Ban Pa Bong Piang is relatively steep with an elevation of 800 to 1400 m above sea level (m a.s.l.). Most of the farmers are Karen people who mainly cultivate upland rice and terraced paddy fields once a year. Cultivation begins in June for land preparation, and then rice planting is in July. Harvest occurs around mid-October to the beginning of November, and then the land is left fallow from December to May.

In this study, the similarities of original geography, soil formation and microclimate conditions were well considered for site selection. A terraced paddy field located upstream of the natural water canal for cultivation was selected (18◦32002.800 N, 98◦26048.800 E). This terraced paddy field (level terrace type) was constructed since 1982 (40 years ago), and the terrace risers or walls were built of soil. Selecting this terraced paddy field avoided nutrient and sediment discharges from other terraced upland rice fields. Its elevation ranged between 975 and 1014 m a.s.l. with slopes of 5–17% (Figure 1b). The widths of the flat section of the terrace ranged from 3 to 5 m, and the heights of the terrace riser or wall were between 0.5 and 0.8 m.

Meanwhile, an upland rice field (non-terraced) was selected (18◦32022.200 N, 98◦26034.700 E) as the comparison field, with an elevation of 908–935 m a.s.l. and 19–26% slopes. Upland rice is as traditional rice cultivation in this area, which has been cultivated for more than 70 years. This field has only grown upland rice once a year, whereas most fields comprise the upland rice–maize system (Figure 1c). Meanwhile, an upland rice field (non-terraced) was selected (18°32′22.2″ N, 98°26′34.7″ E) as the comparison field, with an elevation of 908–935 m a.s.l. and 19–26% slopes. Upland rice is as traditional rice cultivation in this area, which has been cultivated for more than 70 years. This field has only grown upland rice once a year, whereas most fields comprise the upland rice–maize system (Figure 1c).

flat section of the terrace ranged from 3 to 5 m, and the heights of the terrace riser or wall

*Agronomy* **2022**, *12*, x FOR PEER REVIEW 3 of 21

were between 0.5 and 0.8 m.

(**b**)

**Figure 1.** *Cont*.

(**c**)

**Figure 1.** Study area. (**a**) Overall study area, (**b**) terraced paddy field and (**c**) upland rice. *2.2. Management Practices*  **Figure 1.** Study area. (**a**) Overall study area, (**b**) terraced paddy field and (**c**) upland rice. The aerial images were taken from Google maps on 25 February 2020. The photos were taken on 14 November 2020 by Noppol Arunrat.

#### Farm management practices were recorded from farm owners in November 2020. At *2.2. Management Practices*

the terraced paddy field, the transplanting method was used for rice cultivation. The nursery field was prepared in May. Rice seeds were placed into bags and soaked in water for 24 h, and then rice seeds in bags were drained and dried for 24 h in a shady area until small roots appeared at the end of the seeds. Then, pregerminating rice seeds were sown by hand in the nursery field. In June, puddling was prepared by a 15-horsepower tractor for all terraced paddy fields. The rice plants were removed from the nursery field and transplanted in terraced paddy fields by hand. The N, P2O5 and K2O chemical fertilizers were applied using 46-0-0 (93.8 kg·ha−1), 16-16-8 (156.3 kg·ha−1), and 16-20-0 (156.3 kg·ha−1). Harvesting was performed in October by hand. All rice residues were left in the field without burning. Upland rice cultivation began in May with land preparation including removing Farm management practices were recorded from farm owners in November 2020. At the terraced paddy field, the transplanting method was used for rice cultivation. The nursery field was prepared in May. Rice seeds were placed into bags and soaked in water for 24 h, and then rice seeds in bags were drained and dried for 24 h in a shady area until small roots appeared at the end of the seeds. Then, pregerminating rice seeds were sown by hand in the nursery field. In June, puddling was prepared by a 15-horsepower tractor for all terraced paddy fields. The rice plants were removed from the nursery field and transplanted in terraced paddy fields by hand. The N, P2O<sup>5</sup> and K2O chemical fertilizers were applied using 46-0-0 (93.8 kg·ha−<sup>1</sup> ), 16-16-8 (156.3 kg·ha−<sup>1</sup> ), and 16-20-0 (156.3 kg·ha−<sup>1</sup> ). Harvesting was performed in October by hand. All rice residues were left in the field without burning.

weeds and vegetation. The no-tillage method was used due to the difficulty of using any machines on the hillslopes, whereas the drilling method was usually applied for planting. A hoe was used to dig the soil at a 10 cm depth, and then rice seeds were dropped by hand. The N, P2O5 and K2O chemical fertilizers were applied, consisting of 16-20-0 (156.3 kg·ha−1) and 16-16-8 (156.3 kg·ha−1). Harvesting was performed in November by hand, and all rice residues were left in the field without burning. *2.3. Soil Sample Collection and Laboratory Analysis*  Upland rice cultivation began in May with land preparation including removing weeds and vegetation. The no-tillage method was used due to the difficulty of using any machines on the hillslopes, whereas the drilling method was usually applied for planting. A hoe was used to dig the soil at a 10 cm depth, and then rice seeds were dropped by hand. The N, P2O<sup>5</sup> and K2O chemical fertilizers were applied, consisting of 16-20-0 (156.3 kg·ha−<sup>1</sup> ) and 16-16-8 (156.3 kg·ha−<sup>1</sup> ). Harvesting was performed in November by hand, and all rice residues were left in the field without burning.

#### Topsoil was collected at depth 0–10 cm on terraced paddy and upland rice fields after *2.3. Soil Sample Collection and Laboratory Analysis*

harvest in November 2020. At each field, three transects were designed with a distance of 7–8 m from one transect to another; then, the soil samples were collected from each point of each transect. At the terraced paddy field, the uppermost toposequence point was called T1, followed by T2, T3, …, and T24 (lowermost toposequence point), respectively (Figure 1b). At each point of each transect, the soil sample was collected using a mini shovel, then 1 kg of soil sample was packed into a plastic bag. Thus, 72 soil samples were obtained from the terraced paddy field. At the upland rice field, the uppermost toposequence point was called U1, followed by U2, U3…, and U25 (lowermost toposequence point), respectively (Figure 1c). A total of 75 soil samples were gathered from the upland rice field. A compass was used to identify the slope gradient of each soil sample point. This is because the slope gradient influences the rate of runoff on the soil's Topsoil was collected at depth 0–10 cm on terraced paddy and upland rice fields after harvest in November 2020. At each field, three transects were designed with a distance of 7–8 m from one transect to another; then, the soil samples were collected from each point of each transect. At the terraced paddy field, the uppermost toposequence point was called T1, followed by T2, T3, . . . , and T24 (lowermost toposequence point), respectively (Figure 1b). At each point of each transect, the soil sample was collected using a mini shovel, then 1 kg of soil sample was packed into a plastic bag. Thus, 72 soil samples were obtained from the terraced paddy field. At the upland rice field, the uppermost toposequence point was called U1, followed by U2, U3 . . . , and U25 (lowermost toposequence point), respectively (Figure 1c). A total of 75 soil samples were gathered from the upland rice field. A compass was used to identify the slope gradient of each soil sample point. This is because the slope

gradient influences the rate of runoff on the soil's surface, soil erosion and the movement of nutrients. Steel soil cores (5.0 cm width × 5.5 cm length) were used to collect soil samples from both fields for soil bulk density analysis.

At the laboratory, soil bulk density was measured by the dry weight per volume of soil core after drying in an oven at 105 ◦C for 24 h. All soil samples were air-dried at room temperature for 7–14 days, and then the dry soils were crushed and passed through a 2 mm sieve. Soil texture was determined by a hydrometer. Electrical conductivity (ECe) in saturation paste extracts was measured by an EC meter [44]. Soil pH was determined in a 1:2.5 soil to water mixture by a pH meter [45]. The cation exchange capacity (CEC) was determined by the NH4OAc pH 7.0 method. Organic carbon (OC) was determined following the method described by Walkley and Black [46] using potassium dichromate (K2Cr2O7) in sulfuric acid and reported as organic matter (OM) by multiplying with 1.724. Ammonium nitrogen (NH4-N) and nitrate–nitrogen (NO3-N) were measured by the KCL extraction method. Available phosphorus was determined based on the molybdate blue method (Bray II extraction) [47]. Available potassium, calcium and magnesium were extracted by NH4OAc pH 7.0 and measured by atomic absorption spectrometry [48].

## *2.4. Soil Organic Carbon Calculation*

The SOC stock was calculated using the following equation:

$$\text{SOC stock } = \text{OC} \times BD \times L \tag{1}$$

where SOC is the soil organic carbon stock (Mg·C·ha-1), *OC* is organic carbon (%), *BD* is bulk density (g·cm−<sup>3</sup> ) of soil and *L* is soil thickness (10 cm for this study).

#### *2.5. Soil Erodibility Calculation*

Among the estimators of the soil erodibility (*K*-value), the nomograph by Wischmeier et al. [49], the *K*-value in the Environmental Policy Integrated Climate (EPIC) model [50] and a formula proposed by Shirazi and Boersma [51] were widely used. Although there is no study in Thailand, several studies in China (e.g., Du et al. [52]; Wu et al. [53]; Chen et al. [54]; and Liu et al. [55]) achieved scientific results and revealed that the method of the EPIC model is reasonable to calculate *K*-value. Therefore, the *K*value in our study area was calculated by following the EPIC model equation developed by Williams et al. [50] and Sharpley and Williams [56]:

$$K = \left\{ 0.2 + 0.3 \exp\left[ -0.256 \text{SA} (1 - 0.01 \text{SA}) \right] \right\} \left( \frac{\text{SI}}{\text{SI} + \text{CL}} \right)^{0.3} \left[ 1 - \frac{0.25 \text{C}}{\text{C} + \exp(3.72 - 2.95 \text{C})} \right] \tag{2}$$

$$\times \left\{ 1 - \frac{0.7 \left( \frac{1 - \text{SA}}{100} \right)}{\left( \frac{1 - \text{SA}}{100} \right) + \exp\left[ -5.51 + 22.9 \left( \frac{1 - \text{SA}}{100} \right) \right]} \right\} \times 0.1317$$

where SA is the fraction of sand (%), SI is the fraction of silt (%), CL is the fraction of clay (%), C refers to the organic carbon content (%) and 0.1317 is the conversion factor for United States business units (t acre h/100 acre/ft/tanf/in) to the international system of units. Thus, the unit of *<sup>K</sup>* value is t·h·MJ−<sup>1</sup> ·mm−<sup>1</sup> (metric ton (t) × hour (h)/megajoule (MJ) × millimeter (mm)).

#### *2.6. Statistical Analysis*

The differences in soil properties between the terraced paddy and upland rice fields were compared by using independent *t*-tests. One-way ANOVA was employed to test for the differences in soil parameters within the group. The correlation between soil properties in terraced paddy samples and upland rice samples was separately investigated by Pearson's correlation coefficient. Meanwhile, principal components analysis (PCA) was applied to analyze the factors that influence the SOC stocks and soil erodibility under the terrace and upland rice paddies. The variables with high collinearity, correlation coefficients |r| > 0.7, were cut off in Pearson's correlation analysis. These statistical

analyses and visualization were performed using R environment (v.4.0.2) with packages such as 'agricolae' [57], 'ggplot2' [58] and 'factoextra' [59]. Moreover, trend analysis was performed to analyze the variation of soil properties along the slope gradient using SPSS (v. 20.0). pH 5.36 0.29 1.45 0.23 5.31 0.20 5.06 0.03 −1.15 0.25 Bulk density (g·cm<sup>−</sup>3) 1.46 0.03 4.60 0.04 1.35 0.01 3.43 0.07 −32.24 <0.01 Organic matter (%) 2.55 0.33 0.33 0.57 2.79 0.85 128.70 <0.01 2.29 0.02 Soil organic carbon (Mg·C·ha<sup>−</sup>1) 21.61 2.72 0.79 0.38 21.84 6.52 132.20 <0.01 0.28 0.78

**Table 1.** Independent *t*-test comparing soil properties between upland rice (n = 75) and terraced

**Mean Std F** *sig.* **Mean Std F** *sig. t* **Value** *p* **Value** 

not always increase soil nutrient status, as found in the present study, as most soil nutrients (NH4-N, NO3-N, available K, available Ca and available Mg) in the terraced paddy field were lower than the upland rice field (Table 1). Probably, the continuous water flows from plot-to-plot until downstream locations cause dissolved nutrients to be washed and removed from the flat surface (cultivation section), indicating a short period for accumulating nutrients into the soil that results in lower remaining nutrients in the soil. The draining of water at least 7 days before harvesting is another hypothesized cause for the reduction in soil nutrients. Schmitter et al. [63] suggested that soil fertility in rice paddy terraces can be maintained by balancing sediment inputs of different sources. However, continuous observation is needed to estimate any significant trends of these two cropping systems

#### **3. Results and Discussion** Electrical conductivity (dS·m<sup>−</sup>1) 0.24 0.05 0.06 0.81 0.29 0.06 6.08 0.02 5.32 <0.01

paddy field samples (n = 72).

in our study area.

#### *3.1. Variation of Soil Properties under Terraced Paddy Field and Upland Rice* Cation exchange capacity (meq·100 g<sup>−</sup>1) 13.87 1.75 0.07 0.79 8.50 1.45 145.70 <0.01 −20.21 <0.01

**Soil Properties Upland Rice Terraced Paddy Field Between Group**

*Agronomy* **2022**, *12*, x FOR PEER REVIEW 7 of 21

The proportion of topsoil textural composition of each sample of terraced paddy field and upland rice is shown in Figure 2. In the terraced paddy field, topsoils predominantly contained a high proportion of sand particles (49.3–67.1%), which varied significantly among different toposequences (*p* < 0.01, Table 1) from sandy loam to loam. Meanwhile, 29.1–47.9% of the sand proportions of topsoils varied from sandy clay loam to clayey loam in the upland rice field. Similarly, silt and clay contents were significantly different between the terraced paddy and upland rice fields (*p* < 0.01, Table 1), and higher average values were found under upland rice compared to terraced paddy fields (Table 1). In both cropping systems, there was a large difference in the particle size distribution depending on toposequence, as shown in Figure 3. NH4-N (mg·kg<sup>−</sup>1) 29.05 7.36 59.87 <0.01 22.85 6.68 0.59 0.45 −5.34 <0.01 NO3-N (mg·kg<sup>−</sup>1) 17.44 4.52 14.16 <0.01 14.81 4.03 1.24 0.27 −3.72 <0.01 Available P (mg·kg<sup>−</sup>1) 2.57 0.97 8.83 <0.01 95.83 37.49 91.58 <0.01 21.54 <0.01 Available K (mg·kg<sup>−</sup>1) 215.25 48.84 23.97 <0.01 48.00 15.74 4.50 0.04 −27.71 <0.01 Available Ca (mg·kg<sup>−</sup>1) 1063.32 341.00 2.99 0.09 652.63 188.83 46.37 <0.01 −8.98 <0.01 Available Mg (mg·kg<sup>−</sup>1) 141.62 27.97 28.59 <0.01 28.28 7.52 5.85 0.02 −33.25 <0.01 %Sand 41.50 4.52 2.35 0.13 55.70 4.77 0.004 0.95 18.53 <0.01 %Silt 28.96 1.41 2.08 0.15 23.20 3.00 18.86 <0.01 −14.99 <0.01 %Clay 29.54 5.04 3.22 0.08 21.10 3.90 10.65 <0.01 −11.33 <0.01 Soil erodibility (t·h·MJ<sup>−</sup>1·mm−1) 0.2402 0.0136 0.00 0.99 0.2351 0.0104 35.08 <0.01 0.17 0.86

**Figure 2. Figure 2.** Particle size distribution for terrace Particle size distribution for terraced paddy field and upland rice areas. d paddy field and upland rice areas.

From the independent *t*-test in Table 1, the terrace paddy soils showed significantly higher values (*p* < 0.05) of OM, ECe, available P and sand fraction than the upland rice soils. In addition, there were significantly lower (*p <* 0.01) values for bulk density, CEC, NH4-N, NO3-N, available K, available Ca, available Mg, silt fraction and clay fraction in terraced paddy than upland rice soils. However, there was no significant difference in pH, SOC and soil erodibility between these two cropping systems (*p* > 0.05).

In the same cropping system, more variation in soil properties for each toposequence was mostly found in terraced paddy soils rather than upland rice soils. Soil pH, OM, SOC, ECe, CEC, available P, available Ca, available Mg, silt and clay contents as well as soil erodibility significantly fluctuated in different toposequence patches (*p* < 0.05) under terraced paddy cultivation (Table 1 and Figure 3). Meanwhile, significant differences (*p* < 0.05) of bulk density, NH4-N, NO3-N, available P, available K and available Mg were

detected under the upland rice cultivation area (Table 1 and Figure 3). It is noteworthy that the available soil base elements responsively fluctuated under both upland rice and terrace paddy fields in different toposequences, especially available Ca (varied in ranges of 590.9–2302.2 mg·kg−<sup>1</sup> and 198.7–984.2 mg·kg−<sup>1</sup> , respectively, Figure 3).

**Table 1.** Independent *t*-test comparing soil properties between upland rice (*n* = 75) and terraced paddy field samples (*n* = 72).


The relative fluctuations of soil physical and chemical properties were reflected in the impact of patches heterogeneity and soil erodibility as well as practice management in these two areas. The fluctuation of soil physical and chemical properties in the terraced paddy field indicated the high variation of soil characteristics among plots from the uppermost to lowermost toposequence plots. Cui et al. [60] also found that the available K and available P were low in the fallow period compared with the tillage period, whereas OM was higher in the fallow period than that in the tillage period. *Agronomy* **2022**, *12*, x FOR PEER REVIEW 8 of 22

 **Figure 3.** *Cont*.

**Figure 3.** Comparison of soil physical and chemical properties between terraced paddy field and upland rice. **Figure 3.** Comparison of soil physical and chemical properties between terraced paddy field and upland rice.

Based on trend analysis (Figure 4), there were no significant trends of soil physical and chemical properties with the topographic slopes. It is implied that the slope gradients did not influence to the variation of soil physical and chemical properties of both upland rice and terraced paddy fields. The trend line of the terraced paddy field showed that OM, silt content, pH, ECe, CEC, NH4-N, available Ca, available K and available P decreased from the uppermost toposequence downwards the slope. Meanwhile, the trend line of upland rice field indicated that bulk density, OM, silt and clay contents, available K and available P decreased from the top slope to the lowermost toposequence point. This is mainly due to the transportation of nutrients and sediments from upstream to downstream locations under the plot-to-plot irrigation system, which detected the significant differences (*p* < 0.01) of silt and clay particles among terraced paddy soils (Table 1). This is consistent with Miyamoto et al. [61] and Mori et al. [62], who observed that fine particles in the surface were washed and transported downstream by surface flow. Terracing does not always increase soil nutrient status, as found in the present study, as most soil nutrients (NH4-N, NO3-N, available K, available Ca and available Mg) in the terraced paddy field were lower than the upland rice field (Table 1). Probably, the continuous water flows from plot-to-plot until downstream locations cause dissolved nutrients to be washed and removed from the flat surface (cultivation section), indicating a short period for accumulating nutrients into the soil that results in lower remaining nutrients in the soil. The draining of water at least 7 days before harvesting is another hypothesized cause for the reduction in soil nutrients. Schmitter et al. [63] suggested that soil fertility in rice paddy terraces can be maintained by balancing sediment inputs of different sources. However, continuous observation is needed to estimate any significant trends of these two cropping systems in our study area. *Agronomy* **2022**, *12*, x FOR PEER REVIEW 9 of 21  **Figure 3.** Comparison of soil physical and chemical properties between terraced paddy field and upland rice.

**Figure 4.** *Cont*.

**Figure 4.** *Cont*.

**Figure 4.** Trend analysis charts of soil physical and chemical properties between terraced paddy field and upland rice. **Figure 4.** Trend analysis charts of soil physical and chemical properties between terraced paddy field and upland rice.

#### *3.2. Soil Organic Carbon Stock under Terraced Paddy Field and Upland Rice 3.2. Soil Organic Carbon Stock under Terraced Paddy Field and Upland Rice*

The OM fraction was significantly higher (*p* < 0.01) in terraced paddy soils (2.8 ± 0.9%) than in the upland rice soils (2.6 ± 0.3%) (Table 1). There was no significant difference in SOC stocks in both cropping systems, which was because the combination of OM and bulk density variations caused insignificant differences of SOC stocks between both systems. In the upland rice system, the lowest SOC stock (15.0 ± 0.2 Mg·C·ha−1) was found in U1, which is the uppermost toposequence plot (Figure 4). Meanwhile, the sample soil at U12 had the greatest SOC content (26.7 ± 0.1 Mg·C·ha−1) (Figure 5a). However, significant variations of SOC among different toposequences were not detected under upland rice soils. In the terraced paddy field, conversely, SOC stock shows significant differences among toposequence patches (*p* < 0.01, Table 1). It was found that T1, which is the uppermost toposequence plot, contained the largest SOC stock (40.2 ± 0.4 Mg·C·ha−1) (Figure 5b). Meanwhile, the lowest SOC stock was found in T24 (9.3 ± 0.2 Mg·C·ha−1), which was located at the lowest toposequence (Figure 5b). This can be explained by previous evidence of the effect of toposequence on soil properties. Baskan et al. [64] revealed that soil properties in different topographic positions were significantly different in terms of pedogenic processes shaped by the movement and transport of soil particles. In this study, a higher level of SOC content in upper terraces compared to the upland site indicates its potential on SOC sequestration from the above position. De Blécourt et al. [65] quantified The OM fraction was significantly higher (*p* < 0.01) in terraced paddy soils (2.8 ± 0.9%) than in the upland rice soils (2.6 ± 0.3%) (Table 1). There was no significant difference in SOC stocks in both cropping systems, which was because the combination of OM and bulk density variations caused insignificant differences of SOC stocks between both systems. In the upland rice system, the lowest SOC stock (15.0 <sup>±</sup> 0.2 Mg·C·ha−<sup>1</sup> ) was found in U1, which is the uppermost toposequence plot (Figure 4). Meanwhile, the sample soil at U12 had the greatest SOC content (26.7 <sup>±</sup> 0.1 Mg·C·ha−<sup>1</sup> ) (Figure 5a). However, significant variations of SOC among different toposequences were not detected under upland rice soils. In the terraced paddy field, conversely, SOC stock shows significant differences among toposequence patches (*p* < 0.01, Table 1). It was found that T1, which is the uppermost toposequence plot, contained the largest SOC stock (40.2 <sup>±</sup> 0.4 Mg·C·ha−<sup>1</sup> ) (Figure 5b). Meanwhile, the lowest SOC stock was found in T24 (9.3 <sup>±</sup> 0.2 Mg·C·ha−<sup>1</sup> ), which was located at the lowest toposequence (Figure 5b). This can be explained by previous evidence of the effect of toposequence on soil properties. Baskan et al. [64] revealed that soil properties in different topographic positions were significantly different in terms of pedogenic processes shaped by the movement and transport of soil particles. In this study, a higher level of SOC content in upper terraces compared to the upland site indicates its potential on SOC sequestration from the above position. De Blécourt et al. [65] quantified SOC stock changes induced by terrace construction in the area of forest conversion to rubber plantation. They found that topsoil removal at the cut section caused a reduction in SOC stocks in the youngest plantation, and then the recovery of SOC stocks came from roots and litter and deposition of topsoil materials from the upper slope.
