The study area is located in Yuzhong County, Gansu Province, China (36°02′N, 104°25′E, 2,400 m) ( Figure 1 ). The average annual temperature is 6.5°C, and the monthly average maximum and minimum temperatures are 19.0°C and −8.0°C, respectively. The mean annual precipitation from 2000 to 2018 was 323.5 mm, the maximum year was 2018 (458.8 mm), and the minimum precipitation occurred in 2004 (202.0 mm). The average annual water evaporation is approximately 1,450 mm. The local soil type is Calcic Kastanozems (Siltic) or rusty dark loess soil with a pH of 8.2 (Guo et al., 2010). The rainfall data were derived from the perennial positioning automatic meteorological recording instrument ( Figure 1 ).

We selected five alfalfa grasslands in October 2018 and formed a time series of 5, 8, 10, 15, and 20 years, which represented successional series with similar topographic conditions. Where fields planted with alfalfa were recorded in 2013, 2010, 2008, 2003, and 1998, respectively. Abandoned farmland was chosen as the control group (CK; abandoning 15 years). The aspect and slope of each selected alfalfa field remained relatively consistent ( Figure 1 ). In all the selected plots, the alfalfa variety Algonquin was planted, each plot was sown with alfalfa seeds of 35 kg ha⁻¹, the depth was approximately 2 cm, and each row was almost 20 cm. There was no irrigation throughout the growing season. Fertilization was only applied in the first year of planting (N, 185.6 kg ha⁻¹; P, 48 kg ha⁻¹), and no fertilization was conducted afterward. There was no grazing treatment for the alfalfa grassland, which was harvested in June and October of each year (weeds were also mowed at the same time). The control group had no fertilization and mowing treatment. The only difference in the selected alfalfa grassland was in planting years. In the alfalfa grassland, weed biomass increased with prolonged plantation years. The dominant vegetation species are Artemisia frigida Willd., Leymus secalinus (Georgi) Tzvel., Poa annua L., Aster altaicus Willd., Stipa capillata L., and Agropyron cristatum L. Gaertn. The weeds were periodically cleared by hand when needed to avoid interspecific disturbance.

In October 2018, 15 plots of 1 m × 1 m of each field were randomly selected and marked to measure the biomass of each field. Soil samples of 0-cm to 20-cm depth were randomly collected from the field using a soil auger (diameter of 5 cm) with five replicates for each planting period, and each alfalfa field was mixed in three samples. Before air-drying, the soil was sieved through 2.0-mm screens for available nutrient analyses, and, after drying, it was sieved through 0.15-mm screens for total nutrient analyses. The soil aggregate sample was collected with a hard Polyvinyl Chloride (PVC) pipe with a diameter of 10 cm and a height of 20 cm. After sampling, the aggregate sample was divided into small clods a diameter of 1 cm for air-drying.

Soil bulk density (SBD) was determined using cutting ring (0–20 cm). Soil water storage (SWC) was measured gravimetrically (0–20 cm, 20–40 cm, 40–60 cm, 60–80 cm, and 80–100 cm). The soil aggregate-size fraction was measured using a dry- and wet-sieving method (Su et al., 2009). For the dry sieving method, we used 100 g of air-dried soil samples in a stainless-steel vibrating sieve divided into five sizes: >2 mm, 2–1 mm, 1–0.5 mm, 0.5–0.25 mm, and <0.25 mm. The wet sieving method used 50 g of air-dried soil, which was weighed according to dry sieving and moistened for 10 min with distilled water. The apparatus specifications were an oscillation of 10 min at a frequency of 30 cycle min⁻¹. Aggregates retained in the sieve were air dried at 65°C, weighed, and stored for C and N content measurements. The soil aggregate distribution is calculated with Equation 1. The mean weight diameter (MWD) and percentage of aggregate destruction (PAD) are calculated using Equations 2 and 3, respectively.

where AD is the aggregate distribution of each fraction (%), ADi is the i size aggregate weight (g), and TO is the total weight of the soil (g). Xi is the mean diameter of the aggregate, and ADd and ADw are the proportions of dry and water stable aggregates (>0.25 mm), respectively.

The SOC content and C of aggregate fractions are determined by Puget et al. (2000). The soil total nitrogen (TN) content and TN of the aggregate fractions were measured by Bremner (1997). The above- and underground biomass of alfalfa and weeds (soil depth up to 40 cm) were collected from a 1 m × 1 m area of each plot (collected twice in late June and early October). The above- and underground biomass were dried at 105°C for 30 min and then at 65°C to constant weight (Wilson et al., 2009). The C and N contents of alfalfa plants were measured following Yu et al. (2021); aggregate C and N stocks were measured according to Equation 4 (Zhang L. Q.  et al., 2015).

where C and N stocks are stocks of soil carbon and nitrogen [Mg ha⁻¹; where M is soil quality per unit area (Mg ha⁻¹)], SOCi and TNi are different size aggregates of soil carbon and nitrogen content (g kg⁻¹), SBD is the soil bulk density (g cm⁻³), and H is 20 cm.

All statistical data were tested for normality and homogeneity of variance before further analysis. One-way ANOVA was used to compare the differences among the six treatments. The difference was tested by Tukey’s honestly significant difference (HSD) to check the statistical significance of the treatments between different planting years. Mean comparisons were performed using the least significant difference at a probability level of 0.05. Redundancy analysis (RDA) and correlation analysis were used to evaluate the relationship between aggregate C and N stocks, plant biomass, and the C and N concentrations of alfalfa plants. RDA was plotted using Canoco 5.0. Graphs were prepared using ArcGis 10.0 and Origin 2021.

Above- and underground biomass varied significantly among alfalfa cultivation years ( Table 1 ). Throughout various planting for alfalfa, a significant change in above- and underground biomass was occurred. In the aboveground biomass, alfalfa yields reached the highest after approximately 5 years, whereas those of the remaining years declined approximately 20 years after alfalfa planting. Conversely, underground biomass was at a maximum value after alfalfa planting for 8 years but then decreased with the duration years. Therefore, the above- and underground biomass of alfalfa grassland declined with the number of years of planting.

The C and N contents and C-N ratio of alfalfa varied with planting years ( Table 1 ). The C contents of alfalfa above- and underground decreased after continuous planting 5 years. The N content of the aboveground also showed significant change; however, the N content of underground significantly increased with planting years. Similarly, the C-N ratio of alfalfa above- and underground also declined under cultivated 5 years.

As the depth of the soil layer increases, the SWC of the 0-cm to 100-cm soil layer decreases significantly, and the soil moisture content of alfalfa grassland is significantly lower than that of the control ( Figure 2A ). At different planting times, the SWC was the highest after 10 years of planting and the lowest after 8 years of planting. The SBD of alfalfa grassland ranged from 1.04 g cm⁻³ to 1.28 g cm⁻³, and there was a significantly variable trend in 5 years, and there is less change after cultivated 8 to 20 years ( Figure 2B ). The change in SBD after 8, 15, and 20 years of planting was significantly higher than that after 5 years of planting, followed by 10 years of planting.

The soil organic carbon (SOC) content of artificial alfalfa grassland at 0-cm to 20-cm soil depth showed a V-shaped trend with increasing alfalfa planting duration. The highest value of planting for 5 years was 14.4 g kg⁻¹, and the lowest value of planting for 8 years was 8.54 g kg⁻¹ ( Figure 2C ). The soil TN content was between 0.89 g kg⁻¹ and 1.02 g kg⁻¹, and the planting years had a significant effect on soil TN content, which showed a decreasing trend with increasing planting years ( Figure 2D ).

In general, the distribution of soil aggregates in size ranged considerably among various years of cultivation ( Table 2 ). The proportion of aggregates with sizes > 2 mm and < 0.25 mm is significantly higher than that of aggregates with other sizes. After 8 years of planting, the proportion of aggregates > 2-mm size was higher than that for other planting durations, the proportions of aggregates of 1.0–2.0 mm, 0.5–1.0 mm, and 0.25–0.5 mm after 5 years of planting were higher than those for other planting durations (except for the 1.0–2.0 mm). The proportion of <0.25-mm aggregates after 20 years of planting was higher than that for other periods of cultivation.

In addition, both the mechanical and water-stable aggregates, soil aggregate, R_0.25 and MWD showed a declining trend with increase planting years. The mechanical and water-stable aggregates R_0.25 ranged from 62.9% to 43.7% and from 23.7% to 14.2% accordingly. The mechanical and water-stable aggregates MWD change from 1.16 mm to 0.90 mm and from 0.46 mm to 0.35 mm, respectively. In contrast, PAD presented an increasing trend and ranged from 62.2% for planting for 5 years to 67.4% for planting for 20 years.

The soil organic carbon (SOC), TN, and C-N ratios in different aggregates of alfalfa grassland showed a variable trend in planting years. The SOC and TN contents of each size were the highest at 5 years of planting, but the lowest was at 8 years of planting. The C-N ratio in all five sizes showed that the highest value was in the 10-year plantation, and the lowest was in the 5-year plantation (except for the 1–2 mm) ( Figure 3 ).

In addition, the SOC and TN contents and C-N ratio of different size aggregates in the same planting years also showed a significant difference ( Figure 3 ). In general, the SOC and TN contents of macroaggregates in the alfalfa grassland were higher than those in other particle sizes, and the < 0.25-mm particle size was the lowest. In the alfalfa grassland field, the C-N ratio of >2 mm, 0.25–0.5 mm, and <0.25 mm in 8 years was the lowest than that in the other years. After planting for 8 years, the ratios of the five size aggregates increased in all five sizes.

The variation in the C and N stocks of the alfalfa succession series with different planting years is shown in Figure 4 . In general, the C and N stocks of aggregates > 2 mm were the highest in each planting year (except for <0.25 mm). Among different planting years, the stock of C after 5 years of planting was the highest, and, after 8 years, it was declined. Similarly, in each planting year, the N stocks of aggregates > 2 mm were the largest. The N stocks after 5 years were the highest among the five planting years. Overall, the C and N stocks slightly decreased with prolonged planting.

The RDA showed that the CK treatment was concentrated in the first quadrant, and alfalfa grasslands were distributed in the other three quadrants ( Figure 5 ). The first and second axes accounted for 84.63% (a) and 84.62% (b) of the variability explained. Aboveground biomass was negatively correlated with SBD and planting years, whereas underground biomass was positively correlated with planting years. At the same time, we found that aggregate C and N stocks were closely related to the C-N ratio of alfalfa above- and underground plants.

The correlation analysis indicated that aggregate MWD was positively correlated with alfalfa aboveground biomass and the above- and underground C content ( Figure 6 ). The aggregate C stock was markedly positively correlated with the SWC, the C and N content of aggregates, the C content of alfalfa aboveground, and the C-N ratio of above- and underground. Similarly, the aggregate N stock was markedly positively correlated with the SWC, C and N content of the soil surface and each aggregate size, the aboveground C content of alfalfa, and the aboveground C-N ratio. The C and N stocks of aggregates were negatively correlated with underground biomass and the above- and underground N content. In addition, the aboveground C content of alfalfa was positively correlated with SWC and MWD. The above- and underground N content was negatively correlated with the SWC and aboveground C content.

The productivity of alfalfa is affected by soil fertility, soil physical conditions, and other ecological factors (Feng et al., 2022). Soil moisture will affect the growth of roots, and excessive or limited water supply will change the distribution of crop roots, thereby affecting yields (Jia et al., 2009; Fan and Schutze, 2024). The results of this study showed that the largest aboveground biomass was found after 5 years in a cultivated field ( Table 1 ), which is related to the precipitation in previous and subsequent years ( Figure 1 ). Water is the main limiting factor restricting vegetation restoration and reconstruction, and it determines the water–soil ecological relationship of grassland on the Loess Plateau (Jiang et al., 2006). Soil moisture is affected by many factors, such as soil characteristics, land-use structure, topography, vegetation types, and weather, resulting in complex dynamic changes (Fan et al., 2016). In this paper, compared to the control, the soil water content of alfalfa with different planting years has a greater decrease ( Figure 2A ), which indicates that the soil water deficit is obvious. Because the local soil is loess soil, it is beneficial for rainwater to penetrate the soil and restore the water content. On the other hand, we found that the root biomass was negatively correlated with SWC ( Figure 6 ), which may be because the root system of alfalfa is relatively wide, and the proliferation of the root system in the soil increases the porosity of the soil (Jia et al., 2009; Heinze, 2020). In addition, as the planting time increases, the distribution of root systems becomes wider and deeper, which often leads to more consumption and utilization of shallow soil water (Li and Huang, 2008). Moreover, as the planting time increases, the soil moisture gradually decreases, which is directly related to the root biomass. The accumulation of root biomass requires sufficient soil moisture, which leads to water depletion in the soil profile. In addition, alfalfa productivity was also affected by local rainfall (Zhang X. L. et al., 2021). In our study, we found that planting alfalfa for 5 years had the highest aboveground biomass, which may be related to the increasing rainfall from 2010 to 2013, and, after 2013 years, alfalfa biomass showed a declining trend due to the relative decreasing precipitation ( Figure 1 ).

Soil aggregates are closely connected with soil physical, chemical, and biological properties (Plaza et al., 2013; Li et al., 2019). In this experiment, the particle size distribution of soil aggregates in different years of alfalfa cultivation and abandoned land showed a “V”-shaped trend, and soil aggregate size distributions of >2 mm and <0.25 mm were dominant ( Table 2 ). This can be further explained by the fact that the stability of loess soil is generally low. Therefore, macroaggregates will decompose into microaggregates or smaller soil particles after being immersed in water (Lu et al., 2021). In addition, alfalfa planting for numerous years significantly increased the macroaggregate content in certain years, which is directly related to the proliferation of alfalfa roots. Because the root biomass can be increased by producing thick roots or multiple thin roots (Heinze, 2020), the increase in planting time will lead to more accumulation of root biomass, and this extensively proliferating root system helps to consolidate broad soil particles from smaller ones (Tisdall and Oades, 1982). However, due to the reduced physical and mechanical disturbance, the surface herbs, litter, and root biomass increase in natural-restored grassland, resulting in increased soil organic carbon, which is beneficial for aggregate formation (Xiao et al., 2019). Similar outcomes have been reported by Song et al. (2024) who found that the plant residue accelerates the soil aggregates turnover, promoting the soil macroaggregate formation. In addition, we found that dissolved organic matter that, as a substrate for microorganism, contributes to the increase in microbial activity may contribute to macroaggregate formation and stability (Sonsri and Watanabe, 2023).

Plant–soil C and N concentrations and their ratios are essential in the restoration process of grasslands, not only for a better understanding of the C and N cycles but also for management practices (Du and Gao, 2021). We evaluated alfalfa above- and underground plant C and N contents and their ratios during the multiyear growing season. Our results showed that multiple planting durations decreased aboveground plant C and N concentrations and the C-N ratio ( Table 1 ). These findings are supported by years of continuous planting, soil nutrient deficiency, gradual alfalfa grassland degradation, and plant biomass decrease due to self-toxic and self-thinning effects, which eventually results in less photosynthetic function and a lack of C assimilates in alfalfa grasslands (Zhang et al., 2017). At the same time, the results of Table 1 indicated that the N concentration of alfalfa roots increased with increasing planting years, which is similar to the findings of other studies (Cao et al., 2020). The higher N concentrations of roots in plantations than those in natural-restored grassland can be explained by the N₂ fixation of leguminous forage (Song et al., 2021). Additionally, in our paper, we found that alfalfa planting duration resulted in soil C and N decreases in the topsoil layer compared to the control ( Figure 2 ). These results were mainly attributed to decreased plant inputs due to yearly plant removal, a decrease in outside litter input, and increased decomposition rates (Zhang J. H. et al., 2015). Moreover, the degree of C and N decrease depends on many factors, such as basic soil SOC and TN contents, climate conditions, and surface erosion status (Yang et al., 2019).

The SOC content of soil aggregates is a microscopic expression of SOC balance and mineralization rate, which is of great significance to soil fertility and soil carbon sequestration, and the TN content of soil is one of the main indicators of soil fertility (Lu et al., 2021). In this experiment, alfalfa grassland planted for 5 years had higher SOC and TN contents, and the largest SOC and TN storage were found in particles >2 mm and <0.25 mm, whereas the lowest SOC and TN contents were found in particles of 0.25–2.00 mm ( Figure 3 ). This is mainly because macroaggregates are made up of many microaggregates and because of the formation of microaggregates through the combination of organic molecules with clay and cations and not by aggregated particles (Devine et al., 2014). Therefore, soil macroaggregates and microaggregate are rich in organic matter, and soil macroaggregates have a faster turnover time than middle aggregates (Puget et al., 2000). The microaggregates and the surrounding small particles combine to form large aggregates; when the large soil aggregates decompose into microaggregates, the particulate organic matter decomposes, resulting in a greatly reduced SOC content of the microaggregates (Bissonnais and Arrouays, 1997). At the same time, the binding and bonding effects of macroaggregates also reduces the SOC of microaggregates (Tisdall and Oades, 1982). Previous studies have shown that the SOC content of aggregates mainly exists in the particle size of <0.25 mm, which is a combination of smaller organic and inorganic colloids. The smaller aggregates have a larger specific surface area after being combined, resulting in more adsorbed organic matter (He et al., 2011).

Considering the aggregate size distribution and the SOC and TN of specific aggregates, the aggregate size distribution not only represents the contribution rate of aggregates to SOC and TN, but it might also completely depict the interaction of different planting durations to the soil C and N pools (Xu et al., 2012). In different numbers of alfalfa planting years, the stocks of SOC and TN in different proportions of aggregates were mainly restricted by aggregates > 2 mm, which were significantly higher than those of aggregates of other sizes ( Figure 4 ). After 5 years of planting, the SOC and TN reserves of the soil aggregates were the highest. This is because, after 5 years of planting, the soil formed a stable plant community, increased the biomass of underground roots, and accumulated more SOC in the soil. At the same time, alfalfa is a leguminous crop with strong nitrogen fixation ability. The roots fix atmospheric nitrogen, which is conducive to the accumulation of soil nitrogen (Liu et al., 2005; Thomas et al., 2009). In our study, the soil C and N stocks decreased after alfalfa planting duration, which may be caused by the very low C and N inputs from plants and soil. Aminiyan et al. (2015) reported higher aggregate-related SOC content in larger aggregate fractions (> 2 mm). In short, large aggregates > 2 mm are the main contributors to the SOC and TN contents in soil aggregates. This may be due to the increase in the strength or stability of the aggregates due to the humidification of crop residues, thereby increasing the SOC concentration of the large aggregates (Kushwaha et al., 2001). Therefore, increasing the number of aggregates with a particle size greater than 2 mm can enhance soil carbon and nitrogen fixation. These findings may be limited by the method of observing spatial and temporal changes used in this article. However, we will focus on research on long-term positioning to overcome these shortcomings in the future.

Obtaining sustainable high biomass and suitable soil nutrient management is crucial for cultivate alfalfa grassland (Fang et al., 2021); however, degraded continuous alfalfa grasslands are frequently characterized by substantially reduced of biomass output and potential losses in soil nutrient (Gu et al., 2018). In the Loess Plateau of China, the low precipitation and high evapotranspiration of alfalfa field will limit its sustainable development (Ren et al., 2010). The lack of soil water content is the mainly parameter caused the perennial alfalfa grassland degradation (Wang et al., 2019). Alfalfa is a deep root plant that gives it access to water deeper in the soil than annual pastures and crops (Cheng et al., 2005). Alfalfa used soil water from deeper soil profile than other crops and extracted more water and thus creates a large soil water deficit (Huang et al., 2018). Previous study showed that the length of the alfalfa cropping phase in the short term (2–4 years) depends on soil water replenishment, and alfalfa continuously, for 6 years, reduces a relatively desiccation layer in the soil with 2-m to 10-m depth (Li, 1983). Li and Chao (1992) also found that the biomass output of continuous alfalfa markedly decreased after 7 or 8 years due to the declined of soil water. Jia et al. (2009) reported that the alfalfa average yield reached a peak after continuous 9 years planting and, more than 11 years, cultivated had higher soil water use efficiency. In our results, we also found that the soil water content was quickly declined after continuous cultivated alfalfa for 8 years ( Figure 2A ), the dynamic of underground biomass (the underground biomass reached the higher value in 8 year) ( Table 1 ), because the longer the growing years of alfalfa, the deeper its root distribution, which intensified the high water consumption at deep soil (there was no rainfall when we collected soil moisture samples during the first half of the month). On the contrary, the soil water content increased from 10 to 20 years; this was probably because the alfalfa plants’ low transpiration and low productivity made less water consumption. The above results were coincided with the found by Li and Huang (2008) and Ren et al. (2010) and Li et al. (2018). Therefore, the long alfalfa stand duration may deplete available soil water, which would negatively impact production (Li and Huang, 2008) and thus caused the alfalfa grassland degradation.

On the other hand, the deficiency of soil nutrient also limits the alfalfa grassland sustainable development, and eventually leading to alfalfa grassland degradation and, conversely, enhancing the fertilizer application in perennial alfalfa can boost biomass output and prolong the alfalfa degradation timeline (Hakl et al., 2016; Gu et al., 2018; Fang et al., 2021). The decrease of soil nutrients, especially soil nitrogen, limits the perennial alfalfa growth, which, in turn, becomes a crucial restricting factor for the alfalfa cultivation (Yuan, 2017; Fang et al., 2021). In our results, we found that the soil C concentration decreased in the eight year and then increased after the 10th year of alfalfa cultivation ( Figure 2 ). The soil C dynamical dramatically declined mainly because the fertilizer was applied in early planting alfalfa, and the carbon source in the soil is increased. With the continuous cultivated, a large amount of aboveground biomass was removed from alfalfa grassland, and the less carbon source will add to soil. In addition, the root biomass litters were not converted to carbon source by microbe decomposition. However, after cultivation alfalfa for the 10th to 20th years, the soil C storage increased, especially in the microaggregate, and this agrees with the findings from other researchers (Li and Huang, 2008). It seems that soil C storage rose through enhancing the aggregate stability with the increased in cultivation years. However, in this study, TN content continues decreased compared to the fifth year of alfalfa cultivation ( Figure 2 ), mainly because the N that was derived from symbiotic fixation by perennial alfalfa over 20th years does not reach the N value of farmland, and artificial nitrogenous fertilizer losses were larger than N fixation by perennial alfalfa. However, the TN storage in the microaggregate presented less declined and finally increased after continuous cultivation on the 15th to 20th years ( Figure 4 ). This was probably because this growing stage reached an imbalance between TN fixation and loss and reduced the consumption of TN on account of decreased alfalfa plant. Similar results were reported by Ji et al. (2020), who found that the highest values of SOC and TN content were observed in the begin stage of alfalfa grassland, and, with the cultivated prolong, the above contents declined, and the decline soil nutrients restricted alfalfa growth and biomass output.