Introduction

With the population growth and the modernization of agricultural practices, human activities are having an increasing effect on aqueous environments (Hobbie et al. 2017; Tasdighi et al. 2017). For instance, the impact of agricultural production in the USA during the 1960s and 1970s led to a deterioration of the water quality in the Great Lakes, where the proliferation of cyanobacteria seriously affected the survival of fish and other aquatic organisms. Lake Erie, one of these lakes, is now referred to as a “dead sea” in North America (Rudra et al. 2020). A similar situation is also found in China, where roughly 60% of lakes have varying degrees of pollution, 50% of which is caused by non-point source pollution (NPSP) (Zhang et al. 2018). Nitrogen (N) pollution contributes 81% of water pollution in China (Ongley et al. 2010; Yang et al. 2021) conducted a K-means clustering analysis of 253 rainfall events from 2012 to 2018 in De’an County, China and found that intense rainfall was an important factor in local soil erosion, whereas long-duration rainfall had less effect. They also reported that rainfall intensity had a greater effect on soil nitrogen loss than rainfall duration. These studies concur with our results—i.e., rainfall intensity and volume were the main factors affecting nitrogen loss.

Vertical changes in nitrogen loss

The results show that N loss was primarily dissolved-N and nitrate-N in all three vertical spaces, i.e., from rainfall to throughfall to surface runoff, accounting for 72.4–82.1% and 46.2–72.3%, respectively. Accordingly, the proportion of nitrate-N increased gradually on moving from high to low positions vertically. This may be because soil colloids are generally negatively charged and produce more anions (VandeVoort et al. 2013; Bennett et al. 2019). Because the nitrate in nitrate-N is an anion, the repulsion between the two species (soil colloids and nitrate-N) causes the nitrate-N in the soil to be readily soluble in water, as confirmed by Latifah et al. (2017) and Rajta et al. (2020). When rain washes the leaves and surface soil, nitrate-N in the soil dissolves in the water, increasing the proportion of nitrate-N. White et al. (2018) in Australia and Wang et al. (2021) in the purple lands area of southwest China showed that dissolved-N loss accounted for 58.0–70.34% of the total-N loss in runoff, consistent with our findings.

This study has clarified that the concentrations of total-N, dissolved-N, and nitrate-N increased significantly from atmospheric rainfall to throughfall, to surface runoff. In this respect, the increase in the nitrogen concentration in throughfall may have been caused by rainfall scouring and leaching the leaves of plants. Su et al. (2). This suggests that soil nutrient contents could be increased, and soil quality improved by optimizing the stand structure and establishing well-structured water source conservation forests. This would also reduce the risk of water pollution by nitrogen in the watershed (Fig. 4).

Table 4 Effects of soil N content on N loss in surface runoff
Full size table
Fig. 4
figure 4

Effects of different stand structures on throughfall

Full size image

The mechanism of N loss driven by soil and water erosion in water source areas was established through in situ observations based on analyses in the two dimensions of time and space (Fig. 5). From the temporal perspective, the highest N loss (85.9–95.9%) was mainly concentrated in the June–September rainy season, followed by March–May (12.1%). From the spatial perspective (in the vertical direction), N loss gradually increased from atmospheric rainfall (1.5 mg L−1) to throughfall and to surface runoff (2.7 mg L−1). High-intensity rainfall was the main driving force of soil erosion in water source areas.

Fig. 5
figure 5

Mechanisms for the loss of N in water source areas. The size of the circle represents the amount of N loss. TN, total-N

Full size image

Conclusions

Based on our findings, the A. angustifolia population structure differed significantly between protected and unprotected areas. The population structure influenced aboveground biomass and species diversity which were conditioned by protection status. The old-growth populations supported overstocked stands but with lower species diversity. However, lower biomass stock in old-growth populations was also affected by the human disturbance in unprotected forests. The results also highlight the importance of protected areas and sustainable forest management techniques to ensure population renewal and maintenance. Finally, our results indicate that diameter distribution can be a useful guide for sustainable management and for effective protection at the species level.

  1. (1)

    Rainfall intensity and rainfall volume were the key factors affecting nitrogen loss which was mainly in the June–September rainy season, which accounted for 85.9–95.9% of the annual loss.

  2. (2)

    From high to low vertical levels (i.e., from atmospheric rainfall to throughfall and surface runoff), nitrogen concentrations (total-N, dissolved-N, and nitrate-N) gradually increased from 1.5 mg L−1 in atmospheric rainfall to 2.7 mg L−1 in surface runoff. The main forms of loss were nitrate-N and dissolved-N.

  3. (3)

    Optimizing the stand structure of water conservation forests is an effective measure to reduce nitrogen loss in water source areas. For T2, species such as T. chinensis and I. lanceolatum were interplanted and herbaceous plants grew well. Thus, the vertical structure of the stand changed from a mono- to a poly-structure. Tree species with wide crowns captured more throughfall, reducing surface runoff and the migration of nitrogen in this runoff and reduced the total-N loss by 38.2% compared with the pure bamboo stand (T1).

Based on these results, the mechanism of nitrogen loss driven by soil and erosion water is: (1) high-intensity rainfall is the main driving force; (2) nitrogen loss is mainly concentrated in the rainy season (June–September), followed by March–May and October–December, with the lowest loss between January and February; and, (3) the intensity of nitrogen loss increases from high to low vertical positions. A proper stand structure can affect the amount of throughfall, reducing the volume of surface runoff and N losses. This preliminary study provides insights into the mechanism of nitrogen loss in water source areas and contributes to knowledge on reducing losses from forests, thereby hel** to ensure the quality of drinking water and the health of residents is maintained.