在当前全球变暖、大气氮沉降加剧背景下，磷逐渐成为陆地生态系统生产力的主要限制养分。草地作为最大陆地生态系统，在维持区域生态系统稳定和畜牧产品生产中扮演重要的生态和经济角色。然而，由于长期过度放牧导致全球大范围草地发生退化，磷限制日益加剧，这一现象在相对脆弱的干旱半干旱地区尤为明显。因此，研究放牧下草原生态系统磷动态及其内部机制，对于科学放牧管理以平衡草地生态系统稳定发展和畜牧生产需求具有重要意义。本研究以内蒙古温带典型草原生态系统为研究对象，在一个持续7年的放牧强度试验平台开展。放牧强度设置1个对照和三个放牧强度，包括轻度、中度和重度放牧；四个处理下的放牧强度分别为0、0.75、1.50和2.25标准羊单位×天/(公顷×每年)。连续两年对生长季植物和土壤进行采样和监测后，从生态系统到个体水平深入探讨放牧下磷在土壤-植物连续体中的周转过程。研究为揭示温带草原生态系统对放牧和磷限制双重影响的响应机制以及预测未来放牧温带草原生态系统演变动态提供可靠依据。主要研究结果如下：

（1）放牧增加了植被年初级生产力，尤其在降水较多的2020年，其中轻度、中度和重度放牧分别增加了38.9%、116.7%和18.8%，重牧下优势种植物出现退化。放牧加快生态系统磷周转，增加了植物地上磷储量，中牧有最大值283.79 mg m^-2，其主要来源于对土壤磷吸收的增加。轻牧和重牧地上磷储量主要通过增加磷由根系向地上的分配。中牧增加了土壤磷储量，达127.86 g m^-2，而轻牧和重牧降低了土壤磷储量。放牧降低了凋落物磷储量而增加了粪便磷储量，其中中牧下凋落物和粪便分解较快。放牧下地上植被磷供应增加缓解了植物磷限制。

（2）放牧强度对土壤磷有效性有不同影响。轻牧和重牧分别减少了18.2%和30.1%土壤不稳定无机磷(Resin-Pi+NaHCO₃-Pi)，中牧增加了19.0%土壤不稳定无机磷。放牧下不稳定无机磷主要来源于丛枝菌根真菌和放线菌释放的碱性磷酸酶对NaOH-Po的矿化水解，这一过程受土壤pH、含水量以及粉粒比例等影响。放牧下土壤磷组分转化还受凋落物、粪便和根系磷储量以及微生物生物量磷对土壤磷的输入/输出影响。中牧下较高的微生物活性、根系磷储量及利于磷溶解的土壤环境加速了土壤磷循环。

（3）放牧提高优势种羊草和大针茅根际不稳定无机磷(Resin-Pi+NaHCO₃-Pi)浓度，中牧达到最大值，分别为45.61 mg kg^-1和52.76 mg kg^-1。放牧下羊草根际中不稳定无机磷来源于酸性磷酸酶对NaHCO₃-Po和NaOH-Po的矿化；大针茅根际不稳定无机磷则来源于酸性磷酸酶对NaHCO₃-Po和NaOH-Po的矿化，以及低分子量有机酸对NaOH-Pi的动员。根际土壤粒径分布和总氮浓度是影响两物种根际磷组分转化的关键土壤性质。放牧下羊草和大针茅根系共同上调磷酸酶合成及无机磷吸收相关基因，大针茅还上调了三羧酸循环、次生代谢产物合成及遗传物质修复等多个代谢途径基因以维持生长。大针茅更强的生理响应导致的物质和能量消耗是其优势度随放牧强度增加下降更快的原因之一。

（4）放牧增加了羊草和大针茅光合固碳能力，这归因于增加的气孔导度、叶绿素含量、光化学猝灭、电子传递速率、实际光化学量子产量及Rubisco活性等。结合不同磷组分在光合固碳中的功能，羊草通过增加叶片代谢物磷的分配提高光合所需底物浓度，大针茅则通过增加叶片核酸磷的分配提高Rubisco活性并增强叶片对放牧胁迫的抵抗力。羊草更多的叶片代谢相关磷使其具有更高光合能力。放牧增加羊草和大针茅各器官磷和氮含量，但降低各器官碳磷比和氮磷比，加速了磷在生态系统中的循环。

（5）放牧下土壤-植物-微生物在磷动态中密切相关。羊草根际有效磷受土壤磷总量及植物磷吸收影响，大针茅根际有效磷受微生物碳固定及土壤水分条件的影响。两物种植物磷浓度受根际土壤水分、植物氮吸收及根际土壤有效磷的影响。微生物生物量磷受自身碳氮养分利用的影响。综上，放牧下温带草原生态系统磷动态不仅受磷自身的代谢过程影响，其可能更受牲畜活动影响的水分条件以及碳氮养分动态的驱动。

Under the current conditions of global warming and increased atmospheric nitrogen (N) deposition, phosphorus (P) has steadily become the most important limiting nutrient for terrestrial ecosystem productivity. Grassland, as the biggest terrestrial ecosystem, has a crucial ecological and economic function in ensuring regional ecosystem stability and livestock product production. However, P limitation is becoming more severe as a result of global grassland degradation caused by long-term overgrazing, particularly in highly sensitive dry and semi-arid regions. As a result, investigating the P dynamics and internal mechanisms of grazed grassland ecosystems is critical for scientific grazing management in order to balance the stable development of grassland ecosystems with the needs of livestock production. The temperate typical steppe environment of Inner Mongolia served as the research object for this study, which was conducted over a 7-year period on a grazing intensity test platform. Four treatments were set up, consisting of no grazing (control) and three grazing intensities: light, moderate and heavy grazing. The grazing intensity under the four treatments was 0, 0.75, 1.50, and 2.25 sheep units per hectare, respectively. The turnover process of P in the soil-plant continuum under grazing was thoroughly investigated from the ecosystem to the individual level after sampling and monitoring growth season plants and soils for two consecutive years. The study provides a solid foundation for understanding the response mechanism of temperate grassland ecosystems to the simultaneous impacts of grazing and P limitation, as well as forecasting the future evolution dynamics of grazing temperate grassland ecosystems. The following are major study findings:

(1) Grazing increased vegetation's annual primary productivity, particularly in 2020, when precipitation was higher, with light, moderate, and heavy grazing increasing by 38.9%, 116.7%, and 18.8%, respectively, with dominating species degenerating under heavy grazing. Grazing accelerated P cycling in the ecosystem and enhanced plant aboveground P stock. The highest value of grazing was 283.79 mg m^-2, owing primarily to an increase in soil P uptake. Aboveground P stock in light and heavy grazing is primarily accomplished by enhancing the distribution of P from roots to aboveground. Moderate grazing raised soil P stock to 127.86 g m^-2, but light and high grazing decreased soil P stock. Grazing reduced P stock in litter while increasing P stock in manure, while litter and manure decomposed faster under moderate grazing. The increased P supply to aboveground vegetation caused by grazing alleviated plant P limitation.

(2) Different grazing intensities had different effects on soil P availability. Moderate grazing raised soil labile Pi (Resin-Pi+NaHCO₃-Pi) by 19.0% while light and heavy grazing decreased it by 18.2% and 30.1%, respectively. Under grazing, labile Pi is mostly produced via the mineralization and hydrolysis of NaOH-Po by alkaline phosphatase, which is produced by arbuscular mycorrhizal fungi and actinomycetes. Soil pH, water content, and silt proportion all have an impact on this process. The transformation of soil P fractions under grazing was also influenced by P input/output in litter, manure, and root, as well as microbial biomass. Under moderate grazing, the soil P cycle accelerated due to increased microbial activity, root P stock, and a soil condition conducive to P dissolution.

(3) Grazing increased the concentration of labile Pi (Resin-Pi+NaHCO₃-Pi) in the rhizospheres of the dominating species Leymus chinensis and Stipa grandis, with moderate grazing achieving the highest concentrations of 45.61 mg kg^-1 and 52.76 mg kg^-1, respectively. The labile Pi in the rhizosphere of L. chinensis was derived from acid phosphatase mineralization of NaHCO₃-Po and NaOH-Po; the labile Pi in the rhizosphere of S. grandis was derived from the mineralization of NaHCO₃-Po and NaOH-Po by acid phosphatase and mobilization of NaOH-Pi by low-molecular-weight organic acids.

(4) Grazing improved L. chinensis and S. grandis photosynthetic carbon C sequestration capability, which was attributed to increased stomatal conductance, chlorophyll content, photochemical quenching, electron transport rate, actual photochemical quantum yield, and Rubisco activity. Combining the functions of different P fractions in photosynthetic C fixation, L. chinensis increases the concentration of substrates required for photosynthesis by increasing the distribution of leaf metabolite P, and S. grandis increases Rubisco activity and leaf resistance to grazing by increasing the distribution of leaf nucleic acid P. Leymus chinensis has more leaf metabolic-related P, resulting in a better photosynthetic capability. Grazing increased the P and N concentrations of L. chinensis and S. grandis organs, but decreased the C/P and N/P in each organ and accelerated the P cycle in the ecosystem.

(5) P dynamics under grazing are tightly tied to soil-plant-microbe interactions. The overall amount of soil P and plant P intake influenced the Olsen P in the rhizosphere of L. chinensis, while microbial C fixation and soil moisture conditions influenced the Olsen P in the rhizosphere of S. grandis. The two species' plant P concentrations were influenced by rhizosphere soil moisture, plant N intake, and rhizosphere soil Olsen P. Microbial biomass P is influenced by its own C and N nutrient utilization. In conclusion, the P dynamics of temperate grassland ecosystems under grazing are influenced not only by the metabolic process of P, but also by water conditions and C and N nutrient dynamics influenced by livestock activities.