Spatial Distribution Characteristics of Soil C:N:P:K Eco-Stoichiometry of Farmland and Grassland in the Agro-Pastoral Ecotone in Inner Mongolia, China

Abstract

Ecological stoichiometry (ES) is an important index that reflects the balance of various elements in ecological processes. Therefore, it is of great significance to understand the soil nutrient cycle to clarify the environmental control of soil carbon (C), nitrogen (N), phosphorus (P), and potassium (K). In this study, we analyzed the spatial distribution of soil C, N, P, and K contents and the C:N:P:K stoichiometric characteristics of 0–20 cm and 20–40 cm of farmland and grassland in four agro-pastoral areas in Inner Mongolia. Spearman correlation was used to analyze the effects of environmental factors on the soil C:N:P:K stoichiometric relationship. The results showed that there was no fixed Redfield ratio for the soil stoichiometric relationship of farmland and grassland in Inner Mongolia, and the values were 15:2:1:9 to 145:10:1:26 and 25:1:1:29 to 228:15:1:65, respectively. The stoichiometric relationships between farmland and grassland were consistent with the law of geographical and spatial heterogeneity. The ratios of C:N, C:P, C:K, N:P, and N:K showed an N distribution from west to east, while the ratio of P:K showed a V distribution. The stoichiometric relationships in grassland soil were mainly affected by soil organic carbon and total nitrogen content, while those in farmland were mainly affected by total nitrogen and total phosphorus content. The annual mean precipitation has a significant effect on stoichiometric relationships in farmland, while the annual mean temperature has a more significant effect on grassland. In conclusion, the spatial distribution difference in the soil stoichiometric relationship in the agro-pastoral ecotone of Inner Mongolia was more significant than the difference in the land use pattern. The influences of annual mean temperature and annual mean precipitation on soil ecological stoichiometry were in accordance with the geographical spatial similarity law. Compared with grassland, the stoichiometric relationship of farmland soil was greatly affected by fertilization, and farmland in this region was mainly limited by carbon and nitrogen. Thus, field management should be carried out according to local conditions. This study is of great significance as it promotes the rational utilization of land resources and the sustainable development of agriculture.

1. Introduction

Ecological stoichiometry (ES) is a hot topic in current biogeochemical cycle research, which combines the basic principles of chemistry, physics, and biology to study the proportional relationship and the fluxes of various chemical elements in ecological processes [1,2]. In ES research, the Redfield ratio postulates a consistent C:N:P molar ratio of 100:16:1 in marine phytoplankton and open oceanic waters [3]. The C:N:P ratios for all soil layers and organic-rich soil (0–10 cm) in China were 60:5:1 and 134:9:1, respectively [4], and the C:N:P of forests, grasslands, and deserts (0–10 cm) in China was 55:4:1 [5]. However, the law of spatial heterogeneity means that things have spatial differences [6]. Studies in black soil areas in China showed that there were spatial differences in C:N:P stoichiometric characteristics, which are affected by environmental factors [7]. However, the spatial and temporal variation pattern of soil ecological stoichiometry is not completely clear [8]. Do the C:N:P:K stoichiometric relationships of soil follow a certain proportion or the law of spatial heterogeneity at the regional scale in the agro-pastoral ecotone of Inner Mongolia? What is the spatial distribution?

ES is an important predictive index to reflect the composition of organic matter, biogeochemical cycle, and soil quality, and it is also an important index to judge the mineralization and fixation of C, N, P, and K elements [9], which play an active role in ecosystem interactions and represent an important means of understanding the nutrient regulation factors of plant–litter–soil interaction [10,11,12]. The soil C:N ratio (R_CN) is a sensitive index reflecting changes in the soil environment or soil quality. It can be used as an index to measure the mineralization ability of soil C and N, the decomposition rate of organic matter, and the status of nutrient balance. Generally speaking, soil R_CN is inversely proportional to the decomposition rate of organic matter. Soil C:P ratio (R_CP) is a characterization parameter of phosphorus availability. The smaller the soil R_CP, the higher the phosphorus availability in the soil. The N:P (R_NP) in soil is used as an indicator of nutrient-constrained productivity and general biogeochemical status. As a necessary element for plant growth, potassium (k) plays a very important role in the material chemical cycle of the earth, and it has gradually attracted the attention of ecological researchers. In addition, the balance of N, P, and K is beneficial to carbon sequestration in surface soil [1,13]. The study on the spatial distribution of nutrient content and the stoichiometric relationship of C, N, P, and K and their environmental control in the agro-pastoral ecotone of Inner Mongolia has important guiding significance for understanding soil nutrient limitation in each region and adopting reasonable production management measures.

Land use change is a local environmental issue of global importance [14]. Human beings have changed biogeochemical cycles at different scales through farmland reclamation, and the differences in the stoichiometric ratios of C, N, and P in space and time have different impacts on biota [15,16,17]. Located between semi-humid and semi-dry areas, the agro-pastoral ecotone in Internal Mongolia is a transitional belt for farmland and grassland ecosystems and is the most sensitive and unstable area associated with the surrounding environment. From 1947 to 2021, the cultivated land area in the Inner Mongolia Autonomous Region increased from 3,967,000 hm² to 11,567,000 hm² (Statistics Bureau of Inner Mongolia Autonomous Region, 2023), and the change in cultivated land is mainly transformed from woodland and grassland [18]. At present, research on the eco-stoichiometry of soil in the agro-pastoral ecotone of Inner Mongolia is mainly focused on the county level (such as Dalate Banner [19], Yijinhuoluo Banner [20], Duolun County [21], Wengniute Banner [22], etc.) or land types (such as wetlands [23], sandy land [24,25], and grassland [26]), but there are few studies on a large scale in space. However, due to the large longitude span of the agro-pastoral ecotone in Inner Mongolia, there are differences in the topography, soil characteristics, and agricultural planting structures among the agro-pastoral areas.

Previous studies have mainly focused on the C:N:P stoichiometric relationship, while our study innovatively compared the stoichiometry of soil C:N:P:K between farmland and grassland at different scales in the agro-pastoral ecotone of Inner Mongolia. It aimed to clarify the stoichiometric ratio of soil C:N:P:K of different land use patterns and its spatial distribution, analyze soil nutrient deficiency in different regions, and determine the influence of environmental factors on stoichiometry. This provides a theoretical basis that reveals the interaction and balance between C, N, P, and K elements and has important practical significance for understanding the impact of human activities on ecosystem processes and services and exploring agricultural production strategies according to local conditions.

2. Materials and Methods

2.1. Study Site

This study area is the agro-pastoral ecotone from Hulunbuir to Bayannur in Inner Mongolia, China (105°53′–115°31′ E, 40°51′–53°20′ N). According to the Land and Spatial Planning of Inner Mongolia Autonomous Region (2021–2035), the agro-pastoral areas along the Yellow River mainstream plain (I), the foothills of Yinshan Mountain (II), the West Liaohe River Basin (III), and the foothills of Daxing’an Mountains (IV) are, respectively, located from west to east.

The mean annual precipitation in area I is 150–380 mm, and the mean annual temperature is 3.7–7.6 °C. The topography of this area is dominated by plateaus, mountains, hills, and plains, with an elevation of 986–2280 m. Irrigation and silt soil, saline-alkali soil, chestnut soil, and meadow soil are the main soil types, and the main crops are wheat, sunflower, and corn [27]. The mean annual precipitation in area II is 150–500 mm, and the mean annual temperature is 0–6.7 °C. The topography of this area is dominated by plateaus, mountains, and hills, with an elevation of 1150–2350 m. Cinnamon soil and chestnut soil are the main soil types, and the main crops are wheat, rapeseed, sunflower, potato, and miscellaneous grain [28]. The mean annual precipitation in area III is 305–485 mm, and the mean annual temperature is 0–7 °C. The topography of this area is dominated by plateaus, mountains, hills, and plains, with an elevation of 300–2000 m. Soil types include brown soil, black soil, irrigation and silt soil, meadow soil, aeolian sandy soil, and alkaline soil, and the main crops are corn, sunflower, potato, miscellaneous grains, and beans [29]. The mean annual precipitation in area IV is 270–467 mm, and the mean annual temperature is −2–6 °C. The topography of this area is dominated by plateaus, hills, and plains, with an elevation of 150–1800 m. Black soil is the main soil type, and the main crops are corn, soybeans, wheat, and rapeseed [30].

2.2. Data Collection and Sample Analysis

Soil samples were taken from typical farmland and nearby grassland in the agro-pastoral areas of Inner Mongolia during the flourishing period of plant growth. The sampling period of wheat samples in area I was in June, and that of other samples was from August to September. The samples of Hangjinhou Banner, Linhe District, Wulatezhong Banner, Wuchuan County, Wengniute Banner, Keerqin District, Zhalantun City, Yakeshi City, and Labudalin Farm Ranch were sampled in 2019; the samples of Wuyuan County, Dalate Banner, Tumotezuo Banner, Kalaqin Banner, Naiman Banner, Zhalaite Banner, and Sanhe Hui Township were sampled in 2020; and the samples of Taipusi Banner and Chahaeryouyiqian Banner were sampled in 2021. There were 36 farmland sampling sites and 14 grassland sampling sites (Figure 1). In each plot, 4 points with similar terrain and environmental conditions were selected using the diagonal sampling method, and soil samples were collected at 0~20 cm and 20~40 cm soil depths. A total of 288 farmland samples and 112 grassland samples were obtained. Soil samples were air-dried and carefully selected to remove organic matter and fine roots for soil property analysis. Each mixed soil sample was divided into four parts according to the quartering method. One part was selected to pass the 80-mesh sieve to determine the soil pH value and electric conductivity (EC), and the other part was selected to pass the 100-mesh sieve to determine the contents of soil organic carbon (SOC), total nitrogen (TN), total phosphorus (TP), and total potassium (TK). The determination of soil properties follows all standard schemes [31]. Meteorological factors were obtained from the records of 18 national weather stations near each sampling site, and annual mean temperature (MAT), annual mean precipitation (MAP), annual mean relative humidity (MARH), and annual mean sunshine duration (MASD) (1981–2021) were regarded as climate factors. In addition, the data on longitude and latitude for each sampling site were determined using the Global Positioning System (GPS).

2.3. Data Analysis

The ratios of C:N, C:P, C:K, N:P, N:K, and P:K were calculated on a molar mass ratio [32]. The data were tested using a Shapiro–Wilk normal test with SPSS (version 26, IBM SPSS, Somers, NY, USA), and the differences were analyzed via a nonparametric test. The box chart was drawn using Origin (version 2023, OriginLab, Northampton, MA, USA) to show the data on nutrient contents and stoichiometric ratios, and then all the indexes were plotted via Spearman correlation analysis.

3. Results

3.1. Comparison of C, N, P, and K Contents between Farmland and Grassland

Soil nutrients in farmland and grassland were significantly different among agro-pastoral areas (p < 0.05) (Figure 2). On the whole, the contents of SOC and TN in farmland were lower than those in grassland, while the contents of TP and TK were higher than those in grassland, and the difference in TP content in the 0–20 cm soil layer in area I was significant (p < 0.01). Compared with the grassland, SOC, TN, and TK contents in farmland were significantly lower in area I than in other areas, and TP contents were significantly higher than in other areas (p < 0.05). SOC, TN, and TP contents in area III were significantly lower than those in other areas, while TN and TK contents in area IV were significantly higher than those in other areas (p < 0.05).

3.2. Eco-Stoichiometric Characteristics of Soil C:N:P:K in Farmland and Grassland

The differences in the eco-stoichiometric ratios of farmland and grassland (except R_CN) among agro-pastoral areas were significant (p < 0.05) (Figure 3). The change trends of farmland and grassland were the same among areas other than R_PK, while other indexes showed N changes from west to east. On the whole, the R_CN, R_CP, R_NP, and R_NK of farmland were lower than those of grassland, and the R_PK of farmland was higher than that of grassland. The R_CK values of area I and area III were higher than those of grassland, and the opposite was true for area II and area IV. Compared with grassland, the R_CP, R_CK, and R_NK values of farmland in area I were significantly lower than those in other areas; R_PK was significantly higher than that in other areas (p < 0.05); R_CP, R_CK, and R_PK of farmland in area II were significantly higher than those in other areas; R_NP and R_NK of farmland in area III were significantly lower than those in other areas; and R_CN in area IV was significantly higher than that in other areas (p < 0.05).

3.3. Comparison of Nutrient Content and Stoichiometry in the Study Area with Other Scales

Regarding the overall scale of agro-pastoral ecotone in Inner Mongolia, the SOC and TN levels of farmland were lower than those of grassland, but the level of TP was higher, the farmland nutrients were at a middle level, and the level of grassland was at an upper-middle level. The level of TP in farmland in areas I, II, and III was higher than that in grassland, while the levels of SOC and TN in areas II and III were lower than those in grassland. The nutrient level of area III was at the middle and lower levels, and the nutrient level in area IV was the highest (Figure 4).

Compared with the global soil C:N:P, the R_CN of the overall area and four regions in this study was higher, except for the farmland in area III, and the R_CP and R_NP of the farmland in area I and area III and the grassland in all regions were lower. Across the whole study area, compared with the northwest agro-pastoral ecotone, the C:N:P of farmland was higher and the grassland was lower. Compared with the yellow soil area of western Shanxi, the R_CN of farmland was lower, and the other values were higher. Compared with Hebei Province, the R_CN of grassland was lower, and the other values were higher. Compared with Qinghai Province, the R_CP and R_NP of farmland were higher, and the other values were lower. Compared with the study in the same region, the R_CN in area II was lower, and the R_CP and R_NP were higher. Compared with the research in the same region, R_CN in area III was higher and R_CP and R_NP were lower. R_CN and R_NP values in farmland were lower than those in the northern wind and sandy area, and conversely, R_CP in farmland and grassland was higher (Figure 5).

3.4. Relationships between Soil C:N:P:K Stoichiometric and Environmental Factors

Regarding the Spearman correlation between soil stoichiometry and environmental factors (Figure 6), the soil stoichiometry of farmland in area I was significantly correlated with MAP, MARH, and MASD (p < 0.05), while the soil C:N:P of grassland was significantly correlated with MAT and MARH (p < 0.05). The soil stoichiometry of farmland (except R_CK) in area II was significantly correlated with climate factors (p < 0.05), and the soil stoichiometry of grassland was significantly correlated with MAT, MAP, and MARH (p < 0.05). The R_CP and R_NP of farmland in area III were significantly correlated with climate factors (p < 0.01), and the soil stoichiometry of grassland (except R_CN) was significantly correlated with MAT and MASD (p < 0.01), while MASD was opposite compared to other climate factors in this region. In area IV, the R_CN of farmland was significantly correlated with MARH and MASD (p < 0.01); the R_CK was significantly correlated with MASD (p < 0.05); the R_NK was significantly correlated with MAT and MAP (p < 0.05); and the R_CN, R_CK, R_NK, and R_PK of grassland were significantly correlated with climate factors (p < 0.05). In addition, MARH had a different correlation compared to other climate factors in this region. In areas I, II, and III, the correlation between stoichiometry and soil electric conductivity was more significant than pH. The correlations between the contents of C, N, P, and K and the soil stoichiometry of farmland in areas I and II and grassland in area III were significant (p < 0.05).

4. Discussion

4.1. The Spatial Difference in C:N:P:K Stoichiometry Was More Significant than That of Land Use Patterns

On the whole-region scale of the agro-pastoral ecotone, the SOC, TN, R_CN, R_CP, and R_NP of grassland were higher than those of farmland, while TP, TK, and R_PK were lower, indicating that grassland had a stronger nutrient accumulation capacity than farmland. This was similar to the research results of other scholars in agro-pastoral ecozones [21,24,39,40,41]. However, it was inconsistent with previous results in the windy and sandy areas of northern China and the yellow soil area of western Shanxi (Figure 4 and Figure 5), which may be due to the single soil type and low nutrient contents in the study areas. Soil microorganisms in farmland need nutrients to supply their own propagation, and soil available nitrogen content was lower, so R_CN was higher than that in grassland.

Li et al. showed that soil C and N were more sensitive to grassland transformation than P through a meta-analysis of 92 studies [42]. It was believed that crop harvest leads to a reduction in soil C in the farmland ecosystem, that tillage destroys soil structure and accelerates N loss, and that fertilization leads to higher P and K content in farmland. Historically, soils have lost 40–90 Pg C through tillage and disturbance globally [43]. Studies conducted in the agro-pastoral ecotones of northern China [44], northeast China [45,46], and Castelluccio di Norcia (central Italy) [47] areas have also shown that the conversion of grassland to farmland will result in the loss of SOC and a decrease in soil chemical characteristics and basic soil fertility. However, in our study, land use types only had significant effects on TP content, R_CP, and R_PK (p < 0.05), while spatial environmental heterogeneity in the agro-pastoral ecotone of Inner Mongolia had more significant effects on soil nutrient contents and eco-stoichiometric ratios (Figure 2 and Figure 3). This is because soil eco-stoichiometry is jointly regulated by land use patterns, soil properties, human disturbance, climate, and topography factors [48]. The spatial variation coefficient of SOC and TN content in farmland and grassland soil was higher (

Table A2. Descriptive statistics.

Variables	Farmland (N = 144)					Grassland (N = 56)
	Min	Max	Mean	SD	CV(%)	Min	Max	Mean	SD	CV(%)
MAT (°C)	−1.85	8.69	4.61	3.36	72.98	−1.85	7.71	3.96	3.48	87.77
MAP (mm)	135.97	532.33	340.15	97.04	28.53	210.79	532.33	376.39	71.72	19.06
MARH (%)	46.39	68.46	53.46	6.58	12.31	46.39	68.46	55.41	6.35	11.45
MASD (h)	2491.63	3222.56	2902.34	185.65	6.40	2491.63	3098.61	2840.74	168.35	5.93
pH	5.97	8.65	7.70	0.59	7.62	6.15	8.55	7.42	0.73	9.79
EC (μS cm−1)	45.15	989.00	156.53	144.10	92.06	34.35	217.50	90.09	41.94	46.55
SOC (g kg−1)	2.49	43.77	14.24	9.60	67.45	2.01	58.47	21.66	14.79	68.25
TN (g kg−1)	0.27	3.39	1.24	0.70	56.45	0.14	4.10	1.80	1.10	61.15
TP (g kg−1)	0.18	1.18	0.65	0.21	31.82	0.14	1.00	0.52	0.22	41.91
TK (g kg−1)	12.05	27.48	20.56	3.65	17.73	14.05	25.86	20.82	3.14	15.09
RCN	8.06	21.93	13.07	2.42	18.52	8.76	24.85	14.41	3.48	24.17
RCP	12.37	144.82	59.77	35.57	59.51	17.73	231.23	105.90	55.01	51.95
RCK	0.35	5.96	2.22	1.31	58.91	0.26	7.88	3.32	2.06	62.04
RNP	0.86	10.17	4.46	2.22	49.70	1.30	15.49	7.63	3.80	49.81
RNK	0.03	0.39	0.17	0.08	48.81	0.01	0.48	0.24	0.13	54.69
RPK	0.01	0.11	0.04	0.02	46.70	0.01	0.06	0.03	0.01	42.33

), while the variation coefficients of TP and TK were lower because the accumulation of C and N elements was related to the decomposition of organic matter and the self-reproduction ability of soil microorganisms. Therefore, it was greatly influenced by environmental factors. P and K elements were mainly affected by fertilization and soil parent materials, so the variability was small [49].

4.2. Stoichiometric Characteristics Demonstrate Constraints on Agricultural Production in Each Region

The 0–40 cm soil C:N:P:K values of farmland and grassland on the whole scale in the agro-pastoral ecotone of Inner Mongolia were 44:3:1:25 and 82:6:1:36, respectively. The C:N:P of farmland was lower than that of Chinese soil (60:5:1), while that of grassland was higher [4]. The R_CN ratio between 12 and 16 indicated that organic matter was well decomposed. At the same time, a study of forest soil showed that R_CN < 25 indicated a high risk of nitrate leaching. R_CP < 200 indicated net mineralization, and R_NP < 10 represented N limitation [50,51]. The ranges of R_CN, R_CP, and R_NP in this study area were 12.2–15.44, 24.49–145.09, and 2–9.99, respectively, which indicated that soil organic matter in farmland and grassland in this study area could be decomposed well and that the phosphate mineralization rate was high, but nitrogen was limited.

The C:N:P:K values of farmland and grassland in area I were 23:2:1:15 and 37:3:1:24, respectively, being lower than the C:N:P of the national soils. However, the R_CN was higher than the global level, and the R_CP and R_NP were lower, which was similar to the results of Dalat Banner [19]. This indicated that, compared with other research areas, the decomposition rate of soil organic matter in the farmland of area I was slower, and the availability of nitrogen was lower, but the availability of phosphorus was higher [52]. In addition, compared with grassland, the TP content of farmland was two grades higher, and the contents of C, N, and K were at a lower-than-medium level. This may be due to the loss of nitrogen caused by flood irrigation and soil leaching in the Yellow River and the increase in phosphorus caused by over-fertilization, resulting in soil N limitation and P saturation. Therefore, the rational application of phosphorus fertilizer and the management measures to improve the efficient use of soil nitrogen are more conducive to the effective use of agricultural resources in this area, aiding in the achievement of soil element balance.

The soil C:N:P:K values for farmland and grassland in area II were 62:5:1:26 and 100:8:1:36, respectively. The C:N:P values were higher than the national soils, the R_CN was higher than on the global level, and the R_CP and R_NP of grassland were lower. The soil nutrient contents of farmland were at a medium level; C and N were one grade lower than those of grassland, and P was one grade higher. The results indicated that the soil mineralization rate in this area was higher, but it was still limited by a smaller degree of nitrogen. The study in 2019 at the northern foot of Yinshan showed that grassland reclamation changed soil physical structure and increased microbial activity, thereby increasing soil respiration, accelerating the mineralization and decomposition of organic matter, and accelerating the nitrogen loss rate [36]. However, compared with them, the R_CN was lower and the R_NP was higher in our study. This may be because our study included the southern and northern foothills of Yinshan Mountain. The wind erosion intensity of soil in the southern foothills was weaker than that in the northern foothills, and the soil available nitrogen content was generally higher than that in the northern foothills. Therefore, cultivation measures to improve soil carbon sequestration capacity and available nitrogen are more beneficial to the stoichiometric balance of the region.

The soil C:N:P:K values of farmland and grassland in area III were 39:3:1:40 and 87:6:1:81, respectively. The C:N:P of farmland was lower than that of the whole country, while that of grassland was higher. The R_CN of grassland was higher than the global level, the R_CP and R_NP were lower, and the R_CN, R_CP, and R_NP of farmland were lower than the global level. The nutrient content of farmland was lower, and the TP content was one grade higher than that of grassland. The results showed that the decomposition rate of soil organic matter in this area was fast, which was not conducive to the accumulation of organic carbon, and the net mineralization rate of phosphorus was fast, but there was still a strong nitrogen limitation, and the degree of nutrient deficiency was N > C > P. However, the results were at odds with other studies on sandy land, which may be due to the study scale being different [53]. Therefore, reasonable fertilization combined with cultivation measures to improve the ability of soil water and fertilizer retention is beneficial to the material circulation of farmland ecosystems in this region.

The soil C:N:P:K values of farmland and grassland in area IV were 102:7:1:28 and 141:10:1:29, respectively. The C:N:P was higher than the national soil, and the grassland C:N:P (134:9:1) was higher than the national 0–10 cm soil layer. The R_CN, R_CP, and R_NP values of the farmland soil layer were higher than the global level, and the R_CN and R_CP of grassland soil were higher, but the R_NP was lower. The contents of C, N, and K in farmland and grassland soil were at an upper level, and the contents of P were at a medium level. The results indicated that there was no nutrient deficiency in this area, but the net phosphorus mineralization rate was significantly lower. The R_CN of the 0–20 cm soil layer in farmland was higher than that in grassland, which may be due to the low average temperature in the region and the slow decomposition of organic carbon [4,45]. Combined with measures such as crop rotation, fallow, and residue, the surface soil has a strong ability to retain organic carbon.

Early research in Brigelo, Queensland, found that continuous tillage and planting can maintain the availability of soil nitrogen better than continuous grazing [54]. Recent studies in Hokkaido, Japan [55], and Bavaria, Germany, showed that the short-term conversion of grassland to cultivated land increased the diversity of soil bacterial community structure, and combined with the nitrogen fixation of leguminous plants, the organic carbon in the soil increased. Therefore, the dry farming areas in the agro-pastoral ecotone of Inner Mongolia should promote grain–grass rotation, increase the application of organic fertilizer and straw returning to the field, and rationally allocate fertilizer according to the vegetation types and the actual situation in the growth stage, which will help to balance the soil eco-stoichiometry.

4.3. Effects of Environmental Factors on the Eco-Stoichiometry of Farmland and Grassland

Through Spearman correlation analysis, we found that the effects of C, N, P, and K contents on soil eco-stoichiometry in this study area followed the law of geographical correlation [56] (Figure 6). The C content in areas III and IV significantly affected the N, P, and K cycles. K content significantly affected the C, N, and P cycles only in areas I and II. However, N and P content significantly affected the element ecological cycle in the whole region, and R_CN decreased with an increase in P content; therefore, it was not conducive to the accumulation of organic matter. This may be because area I was adjacent to area II, area III was adjacent to area IV, and area I and area III were river basins, so the soil properties between regions were similar. Therefore, the ecological chemical cycle of elements in grassland soil was mainly controlled by C and N, while farmland was mainly controlled by N and P, and R_CN, R_CP, and R_NP were the limiting indexes of soil nutrient content. Similar results had been obtained for the Yellow River Wetland in Baotou [19], Horqin Sandy Land [24], and Chongqing Mountain [57]. In this study, soil pH was only significantly correlated with R_NP, R_NK, and R_PK, which was inconsistent with the research conducted in paddy fields [58] and in the alpine region of the Loess Plateau [59]; however, fertilization caused soil stoichiometry in farmland to be greatly affected by soil pH [60]. This may be related to the land type, geographical environment, and research scale in the study area, so that the conclusions drawn have regional limitations.

The effects of MAT and MAP on soil eco-stoichiometry also had geographical gradients. Soil eco-stoichiometry in areas I and II was affected by MAT and MAP, while MAT and MAP in areas III and IV had the same effects on soil eco-stoichiometry. This may be because areas I and II were located along the Yinshan Mountains in the east–west direction, while areas III and IV were along the Daxing’anling Mountains in the north–south direction. Therefore, the climate and environment of areas I and II and areas III and IV were similar. At the same time, temperature and precipitation were significantly correlated with soil stoichiometry in farmland and grassland, and this was inconsistent with the findings of the northeast black land study [7]. Compared with the mean value of a single year, the average temperature and precipitation in the long-term series of this study can better reflect the long-term climate patterns. Moreover, the influences of temperature and precipitation change on soil stoichiometry in different ecosystems are inconsistent, which contradicts the results of studies on forest and grassland [61]. This may be because farmland is a semi-natural ecosystem and thus different from the natural ecosystem due to human factors.

5. Conclusions

There is no single Redfield-like ratio [3] in farmland and grassland soils across the Inner Mongolia agro-pastoral ecotone, and the R_CN, R_CP, and R_NP of farmland are lower than those of natural grassland. Land use patterns have a significant impact on the cycle of P elements. The stoichiometric relationship between farmland and grassland follows the geographical gradient in the agro-pastoral ecotone, and the trend is consistent. The difference in stoichiometric spatial distribution is more significant than that of the land use pattern.

The deficiency degree of soil nutrients in farmland is C > N, but P is saturated in the agro-pastoral area along the Yellow River mainstream plain, and N > C > P in the agro-pastoral area along the foot of Yinshan Mountain and the West Liaohe River Basin. The effect of agricultural production on the accumulation of C, N, P, and K in soil was not obvious in the agro-pastoral area along the foothills of the Daxing’anling Mountains.

The effect of grassland nutrient content on stoichiometry is related to the similarity of soil properties, while the man-made input of chemical elements destroys the stoichiometric balance of farmland soil, so the stoichiometric relationship of farmland soil is greatly affected by element contents. The effect of MAT and MAP on soil stoichiometry is related to the environmental heterogeneity between regions. Farmland is more significantly affected by MAP, while grassland is more significantly affected by MAT.

In the future, based on this study’s results, the C:N:P:K stoichiometric relationships between soil and plants in farmland can be combined to explore the stoichiometric relationship mechanism in biomass allocation under different agricultural management measures, which can be very useful in improving agricultural production efficiency.