Rapid urbanization has profoundly altered soil carbon cycling and thereby reshaped the spatial pattern of soil carbon content and fractions across the urban-rural gradients. In this study, we measured the contents of total carbon and its different fractions in the topsoil (surface layer 0-10 cm and subsurface layer 10-20 cm) of twenty parks across four urban-rural transects in Beijing, China. We analyzed the spatial variations of different soil carbon fractions and their potential driving factors across the urban-rural gradients. The results showed that topsoil total carbon (topsoil: 21.0±1.6 g/kg; subsurface soil: 18.0±1.3 g/kg) was dominated by organic carbon (topsoil: 64.6%±4.5%; subsurface soil: 54.9%±4.5%). Topsoil contents of organic carbon showed a nonlinear trend from the urban core to the rural area, while the topsoil inorganic carbon content decreased significantly. Topsoil organic carbon (topsoil: 13.8±1.5 g/kg; subsurface soil: 10.0±1.2 g/kg) was dominated by particulate organic carbon (topsoil: 71.3%±2.4%; subsurface soil: 70.5%±2.9%). The contents of both particulate organic carbon and mineral associated organic carbon showed nonlinear changes across the urban-rural forest transects. The proportion of particulate organic carbon was relatively low in urban areas, while that of mineral associated organic carbon showed an opposite trend. Soil texture, soil pH, and park age were important drivers to shape the spatial variation of topsoil carbon components across the urban-rural transects, while the urban-rural climate gradient and species diversity were found to have an unimportant role. Our findings improve the understanding of how urbanization reshapes soil carbon fractions and have useful implications for soil management in urban forests.

Soil moisture is a key driving factor that affects vegetation growth. Vegetation reacts to soil moisture through processes such as evapotranspiration. In-depth exploration of the interaction between soil moisture and vegetation GPP is crucial to ensuring the sustainable development of ecosystems and the efficient utilization of water resources. This study employed MODIS GPP, ERA5-Land soil moisture, and other data sources. It utilized an enhanced nonlinear Granger causality model, along with partial correlation analysis, Sen's slope, and the Mann-Kendall method. These methods were used to examine the combined changes in GPP and soil moisture and their mutual feedback relationship across diverse vegetation types in China from 2000 to 2021. The results showed that: (1) Since 2000, 50.89%-57.61% of the vegetated areas in China have shown a synergistic trend of increasing GPP and decreasing soil moisture, which exhibited a greater proportion with increasing soil depth. The proportion of areas with consistently increasing trends in both GPP and soil moisture was 39.03%-45.76%. (2) In 59.88%-79.38% of vegetated areas, both GPP and soil moisture showed a bidirectional Granger causal relationship. This proportion decreased with soil depth, notably in temperate grassland areas (R6) and temperate desert regions (R7). (3) The increase in GPP resulted in a more significant consumption (57.03%) of soil moisture at 100-289 cm depth and a longer lag effect (2.15 months). Soil moisture at a depth of less than 100 cm mainly promoted the increase in GPP (71.43%-76.58%) and only showed inhibitory effects in some areas, such as the Tianshan Mountains and Hengduan Mountains. The promoting effect of soil moisture at 100-289 cm depth on the vegetation GPP (48.31%) weakened and was accompanied by a significant increase in the lag effect (2.92 months). (4) As the precipitation increased, the interaction between vegetation GPP and soil moisture gradually decreased. When the precipitation was between 200 mm and 400 mm, the interaction was most significant. The influence of different temperature gradients on the interaction between vegetation GPP and soil moisture exhibited multiple threshold effects. This study helps to deepen our understanding of the interaction between carbon and the water cycle of the terrestrial ecosystem in the context of climate change. It also provides an important theoretical reference for the implementation of future ecological restoration projects and the sustainable development of ecosystems.

The deciduous broad-leaved forests are a typical vegetation in the eastern monsoon region of China. This work utilizes the fine classification data of surface cover of composite elevation information to extract the upper limit of montane deciduous broad-leaved forests. We examine the distribution characteristics of the upper limit and its factors influencing the montane deciduous broad-leaved forests by constructing cloud models of the upper limit height. Moreover, this work constructs multiple linear regression models (with the upper limit of deciduous broad-leaved forests at multiple scales (regional, mountain, and local) as the dependent variable and the influencing factors as the independent variables), and a weight coefficient cloud model of influencing factors. Furthermore, this work compares and analyzes the scale changes and spatial differences of the effect of influencing factors on the upper limit of deciduous broad-leaved forests. The sensitivity differences of different montane deciduous broad-leaved forest upper limits to climate factors are also explored. Results show that: (1) The upper limit height of the deciduous broad-leaved forest in the eastern monsoon region of China first increases and then decreases from north to south. The expectation (Ex), entropy (En), and hyper entropy (He) of the distribution height cloud model are 965.77-1993.52 m, 132.80-514.09 m, and 27.58-205.34 m, respectively. (2) Significant scale changes can be observed in the impact mechanism of the upper limit of deciduous broad-leaved forests in the mountainous areas: at the regional scale, the dominant factor for non-climatic and climatic forest lines is mountain base elevation, with contribution rates of 71.36% and 44.06%, respectively. The climatic forest line is more affected by temperature than by precipitation. Meanwhile, non-climatic forest line is more affected by precipitation than by temperature. At the mountain scale, the upper limit of deciduous broad-leaved forests in the mountainous areas is mainly influenced by January average temperature and annual precipitation, and the role of January average temperature in most mountainous areas is larger than that of annual precipitation. On a local scale, except for the Dabie Mountains, the mountaintop effect has the highest weight on the upper limit of deciduous broad-leaved forests in each mountainous area (Ex: 44.84%-68.15%). In addition, the expectation weight of annual precipitation (Ex: 15.45%-41.86%) is higher than that of the January average temperature (Ex: 4.3%-9.97%). (3) The deciduous broad-leaved forests in the Dabie Mountains and Taihang Mountains are most sensitive to annual precipitation (Ex: 40.24% and 18.95%; He: 0.96% and 1.89%). Lvliang Mountains are the most sensitive to January average temperature (Ex: 8.31%; He: 1.09%). Exploring the spatial distribution characteristics and influencing factors of the upper limit of deciduous broad-leaved forests in the mountainous areas can promote the study of differences in altitudinal belt response to climate change and provide theoretical support for the construction and management of regional ecological security monitoring systems.

With the acceleration of global landscape fragmentation, it is of great significance to reveal the impacts of landscape connectivity on biodiversity in order to restore ecological process and protect biodiversity. Taking the grassland of Poyang Lake as research object and from the integrated perspective of combined water level change and species dispersal, this study firstly analyzed the dynamics of grassland structural and functional connectivity respectively after identifying the distribution pattern of grassland at different water levels using landscape pattern indices and the graph-based connectivity indices. Then, the effects of grassland connectivity on plant diversity and the scale effects were discovered by means of the linear regression model. Moreover, the multiple stepwise linear regression model was applied to compare the relative explanatory power of structural and functional connectivity to plant diversity pattern. The results are shown as follows: (1) The grassland in Poyang Lake exhibits a landscape characteristic of "expanding when water falls while shrinking when water rises". It faces both inundation and segmentation effects of water with water level change, and its shrinkage always coexists with landscape fragmentation. (2) With the increase of water level, the area of grassland patch continues to shrink, the shape tends to be simple, the density reduces, and the fragmentation intensifies, resulting in a decrease of grassland structural connectivity. Meanwhile, the number of components increases, the probability of connectivity reduces, so that grassland functional connectivity decreases significantly. (3) The higher the connectivity of grassland, the higher the richness of plant species at the landscape scale, and the higher the similarity of plant communities at the patch scale. Functional connectivity is more effective in explaining plant diversity pattern than structural connectivity does. This study can provide some references for landscape pattern optimization and plant diversity conservation in lake areas from a new perspective of landscape connectivity.