Introduction

Wetlands play a crucial function in protecting water quality of rivers (Gabriel et al. 2008), as wetland soils act as sink, source and transfer of chemical materials (e.g., nitrogen and phosphorus) (Reddy and Delaune 2008). Chen et al. (2012) and Hinsinger (2001) demonstrated that phosphorus could serve as a common limiting element for aquatic plant growth and cause the eutrophication of water bodies. Moreover, soil phosphorus can directly influence the primary production, plant community composition and C and N cycles of wetland ecosystems (Xu et al. 2012). In particular, soil available phosphorus (AP) which can be absorbed by plants is an important indicator to evaluate the capacity of soil phosphorus supply to plant growth (Sun et al. 2012). Seasonal changes in soil phosphorus availability reflect mineralization-immobilization processes of phosphorus and the potential capabilities of phosphorus supply (Chen et al. 2003). Therefore, better understanding dynamic changes in soil phosphorus can contribute to wetland conservation and management.

Phosphorus cycle, especially phosphorus retention and release in wetland soils is greatly impacted by wetland hydrology (Noe and Childers 2007; Gabriel et al. 2008). Song et al. (2007) presented that the drying and re-wetting events in wetlands could cause substantial release of phosphate by chemical changes as the activation of P-mineralization enzyme which could enhance the conversion from organic P to inorganic P (Xiao et al. 2012). With the exception of hydrology, phosphorus fractions in wetland soils were also greatly influenced by other soil properties (Nieminen and Penttila 2004; Kaiserli et al. 2002; Zhang et al. 2011a, b), including soil pH (Sato and Comerford 2005), soil organic matter (SOM) (Liu et al. 2009), salinity (Sundareshwar and Morris 1999) and particle size (Zhang et al. 2011a, b). Some researchers have focused on the biological processes (mineralization-immobilization) and chemical processes (adsorption–desorption and dissolution-precipitation) of phosphorus in grassland, forest and freshwater wetland ecosystems (Chen et al. 2003; Xiao et al. 2012; Noe et al. 2013). However, few studies have been carried out on spatial and temporal distribution of soil P and P availability in a seasonal flooding wetland in the coastal regions (Young and Ross 2001; van Dam et al. 2007).

The primary objectives of this study were (1) to investigate spatial and seasonal distributions of AP and total phosphorous (TP) contents in a seasonal flooding wetland along a sampling zone between the Yellow River channel and a tidal creek in the Yellow River Delta of China; (2) to evaluate the ability of soil phosphorus supply; and (3) to reveal the relationships between soil P and other selected soil properties.

Materials and methods

Study area

The study area is located in the Yellow River Delta of China (N37°45′57.8″E119°11′15.2″; Fig. 1). The Yellow River Delta is one of the youngest wetland ecosystems, and it is also the fastest growing wetland in area among the worldwide larger river deltas (Jiang et al. 2012). It has a semi-humid continental monsoon climate with four seasons, with the annual average temperature of 12.1 °C, and a frost-free period lasting 196 days. The annual average rainfall is 552 mm, and the annual average evaporation is 1,962 mm. The rainfall in this study area mainly concentrates in summer, which often leads to flooding. Meanwhile, the flow-sediment regulation regime in July since 2002 from upstream hydrological station also contributes to flooding during this period from July to October. A large-area floodplain with unique wetland landscape is formed for the accumulation of nutrients deriving from upstream (Cui et al. 2009). The dominant vegetations are Tamarix chinensis, Phragmites australis and Suaeda salsa in the Yellow River Delta (Cui et al. 2008). Generally, soil type in this delta mainly includes coastal saline alluvial soil and marsh soil.

Fig. 1
figure 1

Location map of sampling sites in the Yellow River Delta

Full size image

Sample collection and analysis

A seasonal flooding wetland was selected in the Yellow River Delta. Based on the differences in soil moisture and salinity, four sampling sites (Sites A, B, C and D; Fig. 1) were selected along a 250-m length of sampling zone between the Yellow River channel and a tidal creek in fall (November of 2007) and spring (April of 2008). Soil profiles with three replicates were collected and stratified at depths of 0–10, 10–20, 20–40, 40–60, 60–80, and 80–100 cm. Soils at the same layer were mixed for each sampling site to form a composite soil. A total 48 soil samples were obtained and stored at 4 °C with air-sealed plastic bags in the laboratory before analysis. All soil samples were air-dried and size-fractionated with a nylon sieve (100 meshes) after removing the visible plant litter, coarse root materials, and stones, and then stored in sealed containers before analysis. Another soil core (4.8 cm diameter) for each soil layer of each sampling plot in both seasons were collected for the determination of soil water content and bulk density (BD).

Total phosphorus (TP) and aluminum (Al) were digested in the HClO4-HNO3-HF mixture. The digested solution of soil samples was analyzed by inductively coupled plasma atomic absorption spectrometry (ICP-AAS) (Liu et al. 2012). Quality assurance and quality control were assessed using duplicates, method blanks and standard reference materials (GBW07401) from the Chinese Academy of Measurement Sciences with each batch of samples (1 blank and 1 standard for every 10 samples). The recoveries of samples spiked with standards ranged from 99 to 106 %. Available phosphorus was determined colorimetrically in all the extracts using 0.5 M sodium bicarbonate (Murphy and Riley 1962). Soil moisture content and soil BD were measured by oven-drying at 105 °C for 24 h. Soil particle size was analyzed using a Laser Particle Size Analyzer (Microtrac Inc., USA). Soil organic matter was determined using dichromate oxidation (Walkley and Black 1934). Soil pH and salinity were measured in the suspension of 1:5 soil/water (m/v) using a pH meter and salinity meter, respectively. Total nitrogen (TN) was measured using the Elemental Analyzer (Vario EI, Elementar Co., Germany). The soil properties of these soil samples are listed in Table 1.

Table 1 Physical and chemical characteristics of the sampled soils
Full size table

Total phosphorus stock (TPS) at certain soil layer of each sampling plot in each sampling date was calculated by Eq. (1):

$${\text{TPS}} = {\text{BD}}_{i} \times {\text{TP}}_{i} \times h/100$$
(1)

where TPS (g/m2) is TP stock; BDi (g/cm3) is soil BD of layer i; h (cm) is soil depth; TPi (mg/kg) is TP content at soil layer i (i = 1, 2, 3, 4, 5, and 6).

Statistical analysis

Person correlation analysis was performed to identify the relationship among soil P and selected environmental factors. ANOVA analysis was conducted to test the differences in phosphorus contents among the four sampling plots, and among six soil depths and between two seasons. Differences were considered to be significant if p < 0.05. The stepwise regression analysis was performed to identify the relationship between soil P content and storage and other selected soil properties. Statistical analysis was conducted using SPSS 16.0 software package, and linear graphs and contour maps were conducted using Origin 8.0 and Surfer 10.0 software packages (Xiao et al. 2012; Bai et al. 2012). The inverse distance weighted method was applied to perform contour maps.

Results

Horizontal distributions of AP and TP contents in surface soils in both seasons

The horizontal distribution patterns of TP and AP contents in the top 10 cm soils in the seasonal flooding wetland in both spring and fall were illustrated. As shown in Fig. 2, TP contents in top soils in spring were slightly lower than those in fall (p < 0.05). TP contents in top soils followed the order Sites A (698.6 mg/kg) > Site B (688.0 mg/kg) > Site C (638.8 mg/kg) > Site D (599.2 mg/kg) in fall. TP levels generally increased with increasing distances away from the Yellow River channel. However, TP contents in top soils exhibited an opposite changing tendency in spring, showing the order of Site C (699.6 mg/kg) > Site D (651.7 mg/kg) > Site B (593.6 mg/kg) > Site A (577.5 mg/kg) in spring. Generally, TP contents at Sites A and B were higher in fall, while higher TP contents were found at Sites C and D in spring.

Fig. 2
figure 2

Horizontal distributions of TP and AP in surface soils (0–10 cm) in fall (a) and spring (b)

Full size image

Unlike TP, AP contents in top soils in spring were obviously higher than those in fall. AP demonstrated higher levels at Sites C and D, and lower AP contents at Sites A and B in both seasons. Moreover, the maximum AP level (7.6 mg/kg) appeared at Site C in both seasons. At Sites A and B, higher AP levels in top soils were observed in spring than those in fall, while at Sites C and D, no significant differences in AP levels were observed between fall and spring (p > 0.05).

Vertical distributions of AP and TP contents in both seasons

Figure 3 shows the vertical distributions of TP and AP contents in the seasonal flooding wetland of the Yellow River Delta in both seasons. Generally, TP contents decreased with depth along soil profiles in fall in this sampling zone, except for the patches at the 80–100 cm soil layer of Site A and the 10–20 cm soil layer of Site B. However, TP levels increased with increasing soil depth at Sites A, B and D, whereas a decrease in TP levels was observed at Site C in spring. Compared to fall, TP levels in upper soils decreased at Sites A and B close to the Yellow River in spring, whereas they increased at Sites C and D far away from the Yellow River. However, they increased in deeper soils of Sites A, B and D, while decreased in deeper soils of Site C.

Fig. 3
figure 3

Vertical distributions of TP and AP in 100 cm soil depth in fall (a) and spring (b)

Full size image

The vertical distribution of AP contents did not show significantly variability in the seasonal flooding wetland in both seasons (ranging from 3.1 ± 0.9 to 5.6 ± 2.7 mg/kg in fall and from 2.5 ± 0.7 to 6.2 ± 1.3 mg/kg in spring) (Fig. 3). Similar to TP, except that those obvious patches with lower AP content appeared at the same soil depth of Sites A and B, AP decreased with depth along soil profiles at Sites C and D in fall. Moreover, Sites C and D generally contained higher AP levels than Sites A and B in the whole soil profiles. In spring, AP levels decreased with increasing soil depth in this sampling zone. A large patch with lower AP levels appeared in the bottom soils of Site C. Compared to fall, AP levels were higher in upper soils, whereas they were lower in deeper soils. Moreover, the patches with higher AP levels at Sites A and B in fall were generally diffused among the soil increment of 40–80 cm due to leaching.

As shown in Fig. 4, the mean TP content in this sampling zone exhibited similar profile distributions in both seasons except for surface soils. With the exception of lower levels of TP in the 0–10 and 80–100 cm soil depths, the mean TP level was significantly higher in spring than that in fall, indicating two obvious accumulation peaks at both 20–40 and 60–80 cm soil depths. However, as for the mean AP levels, they showed similar profile distributions in both seasons, with an accumulation peak at the 40–60 cm soil depth. The AP/TP ratios exhibited similar profile distributions to AP in both seasons (Fig. 4).

Fig. 4
figure 4

Profile distributions of TP (a), AP (b), and AP/TP ratios (%) (c) in the seasonal flooding wetland of the Yellow River Delta in fall and spring

Full size image

Profile distribution of soil P stocks in both seasons

Figure 5 shows TPS at each soil layer of each sampling site in both seasons. TP stocks in the 0–100 cm soil depth ranged from 1108.8 to 1146.6 g/m2 at four sampling sites in fall and they didn’t show big differences in TP stocks between different soil layers except for higher accumulation in the bottom soils (80–100 cm) of Site A and the 40–60 cm soil layer of Site B. Comparatively, in spring, the highest TP stocks in the 0–100 cm soil depth were observed at Site D (1065.8 g/m2) among four sites due to higher contribution of TP stocks at the soil layer of 60-100 cm, however, no significant differences in TP stocks were observed among other sites. Although higher TP levels were observed in spring compared to fall, the mean TP stock for each soil layer in this zone was significantly higher in fall (p < 0.05, Figs. 4, 5). Moreover, the mean TP stocks decreased slightly along soil profiles, whereas a big accumulation peak of TP stocks appeared at the 60–80 cm soil depth in both seasons (Fig. 6). The mean TP stock at the 0–100 cm soil depth was 1124.6 g/m2 in fall, which was higher than that in spring (959.9 g/m2).

Fig. 5
figure 5

TP stocks in 100 cm depth in each sampling site in fall (a) and spring (b)

Full size image
Fig. 6
figure 6

Profile distribution of TP stocks in 100 cm depth in the seasonal flooding wetland in fall and spring

Full size image

Relationships between soil P and other selected soil properties

Table 2 lists the Pearson correlation coefficients among TP, AP, TPS and other selected soil properties in the seasonal flooding wetland of the Yellow River Delta. Significantly positive correlations among AP, TP and TPS were observed. AP had significant positive correlations with soil properties such as SOM, salinity, TN, Al, clay, and silt, whereas significantly negative correlations were observed with soil depth and sand. However, TP was just significantly correlated with TN and Al (Table 2). TPS showed significantly positive correlations with BD (p < 0.01), silt and clay (p < 0.05), whereas they were significantly negatively correlated with moisture and pH values (p < 0.01). A significant (p < 0.01) negative correlation between AP and soil depth was observed, but TP was not significantly correlated with soil depth (p > 0.05). The linear regression equations using stepwise method for TP, AP and TPS with combinations of the selected soil properties were given as follows:

$${\text{AP}} = 1.639 {\text{salinity}} + 0.00007{\text{Al}} + 2.892{\text{BD}} - 6. 875 \left( {R^{ 2} = 0. 4 5 6; \,p < 0.00 1} \right)$$
$${\text{TP}} = 0.002 {\text{Al}} + 504.276 \left( {R^{ 2} = 0. 2 6 1;\,p < 0.001} \right)$$
$${\text{TPS}} = 0.017{\text{TP}} + 6.428{\text{BD}} - 10.695 \left( {R^{ 2} = 0. 9 9 7;\,p < 0.001} \right).$$
Table 2 Pearson correlation coefficients among soil P and other selected soil properties
Full size table

The above equations showed that 45.6 % of AP content could be controlled by the salinity, Al and BD, and 26.1 % of TP content could be controlled by Al, whereas TPS was dominantly controlled by TP and BD.

Topsoil concentration factors (TCFs) of TP and AP in both seasons

The topsoil concentration factor (TCF) represents the relative contribution of certain nutrient content at the 0–10 cm depth to the total content in the top 1 meter based on the unit area (Jobbagy and Jackson 2001). Figure 7 showed TCFs for TP and AP in the seasonal flooding wetland of the Yellow River Delta in both fall and spring. TCFs for AP and TP at all four sampling sites in both seasons were generally higher than 0.1, which means that the vertical distribution of phosphorus showed shallower distribution in soil profiles. TCFs for TP was higher at Sites A, B and D than that at Site C in fall, while higher TCF for TP at Site B could be found in spring compared to other sites. Higher TCFs for AP at Sites C and D were observed in fall, whereas they showed higher TCF values at Sites A, C and D in spring. Generally, TCFs for AP are higher at these sites than those for TP in both seasons (except for Sites A and B in fall).

Fig. 7
figure 7

Ranking of TP and AP according to TCFs (relative contribution of the 0–10 cm depth interval to the total in the first meter) from shallow to deep. The horizontal line indicates a topsoil concentration factor of 0.1 corresponding with a homogeneous or random vertical distribution

Full size image

Discussion

Dynamic changes in AP and TP contents in surface soils

Total phosphorus and AP are two main indicators to measure phosphorus levels in soil (Lv et al. 1999). It is important to quantify the temporal distribution of TP and AP in the surface soils as they could be more easily leached by run-off than other soil layers (Gudimov et al. 2011). Higher TP and AP contents were observed in the surface soils of Sites C and D compared to Sites A and B except for TP in fall (Fig. 2), which was closely related to plant P cycles as plant litter could return P to surface soils. Additionally, phosphorus exchanges between water and adjacent wetland soils (Loeb et al. 2008) might be another important reason for lower P contents at these sites adjacent to the Yellow River. This indicates that phosphorus in wetland soils of Sites A and B had a reaction of desorption which would transfer phosphorus from soils to adjacent water. Heredia and Cirelli (2007) found that desorption of phosphorus which bring higher risk of phosphorus from the soils to adjacent water bodies would happen when phosphorus content exceed the soil sorption capacity. A large number of AP would be assimilated by plant roots from the soils during the period from April to September for plant growth, which leads to the transformation of TP to AP. Qin et al. (2007) also presented that AP contents varied with plant growth rhythm. Compared to spring, the cold weather in fall would contribute to P accumulation in the soils, leading to higher TP levels in surface soils (Fig. 2).

AP contents in surface soils were obviously lower in fall than those in spring, which was due to lower soil temperature in fall compared to spring, which limited microbial activities and phosphorus mineralization (Richardson and Simpson 2011). However, higher soil temperature could promote organic phosphorus mineralization (Schmidt et al. 1999), resulting in higher AP levels in spring (Fig. 2). Additionally, the drying and re-wetting cycles caused by the flow and sediment regulation were also one important factor influencing phosphorus distribution in the top soil. (Chepkwony et al. 2001) presented that soil AP increased due to the mineralization of organic matter under the condition of re-wetting after drying. Olila et al. (1997) found that drying condition could increase the readily AP at the expense of labile organic phosphorus in wetland soils. In this study, surface soils contained lower moisture contents in spring than those in fall, which also contributed to higher AP level in spring.

Dynamic changes in AP and TP contents in soil profiles

Higher AP levels generally occurred in upper soils (0–40 cm) (Fig. 3), with the maximum P levels in surface soils (0–10 cm). AP showed significantly negative correlation with soil depth (p < 0.01; Table 2). This is consistent with the result of Chen et al. (2012). However, the profile distribution of AP did not show much higher variability in the seasonal flooding wetland in both seasons compared to Tai Lake region (Chen et al. 2012) and Victoria Lake region (Gikuma-Njuru et al. 2010). In contrast, the profile distribution of TP had no close relation with soil depth due to complicated P forms and soil properties along soil profiles, which is not consistent with the result of Zan et al. (2011). Additionally, soil parent material, soil texture and the stability and mobility of phosphorus also determined the homogeneous distribution of TP (Xiao et al. 2012). Compared to fall, higher TP content and lower TP stocks were observed in soil profiles in spring (Figs. 4, 6). This implies that soil BD was the dominant factor influencing TP stock (Table 2). The BD in spring was relatively lower than that in fall, which indicates that soil was looser and its air permeability and water holding capacity were better in spring compared to fall (Table 1).

Available P can be directly absorbed by plants, which mainly derive from the conversion of TP. Most studies demonstrated that higher AP/TP ratios in soils could promote plant growth (Xiao et al. 2012; Zhao et al. 2009). The AP/TP ratio was also defined as phosphorus activation coefficient (PAC) and considered as an important indicator for soil fertility (Li and Zhang 1994). The AP/TP ratios for all the soils were below the threshold of P load (2 %) for P bioavailability (Xiao et al. 2012) in both seasons in this study, suggesting that the conversion rate from TP to AP was relatively low, and the AP supply for plant growth was deficient. However, TP and AP exhibited similar profile distributions in both seasons, which were mainly controlled by plant root system distribution and soil properties (Stark 1994; Xiao et al. 2012). Meanwhile, plant biological cycle generally moves nutrients upwards by plant assimilations and returns nutrients back to surface soils through plant litter fall (Stark 1994). Therefore, plant P cycles could contribute to the shallower distribution of P in soil profiles. AP generally exhibited shallower distribution compared to TP in the same season in this study as TCFs would increase when nutrient distribution becomes shallower (Jobbagy and Jackson 2001), which indicates that AP was more important for plant growth in this study. Moreover, the higher TCFs of AP at Site C were also associated with these soil properties, such as SOM, pH values, soil texture, etc. in wetland ecosystems. Additionally, the phosphate leaching could explain P accumulation in deeper soils (Figs. 4, 6) as phosphate leaching from soils contributed to transferring AP from topsoil to subsoil (Hesketh and Brookes 2000).

Effects of environmental factors on soil P

It is well known that soil nutrients were closely related to SOM in wetlands (Bai et al. 2007). With the mineralization of SOM, AP will be released and finally increase AP level in soil (Fekri et al. 2011; Tao et al. 2012; Camargo et al. 2013). Additionally, the metabolic activities of microorganisms might also be promoted by SOM and influence phosphorus transformation in soil (Vincent et al. 2010). A significantly positive correlation between AP and SOM was also observed in this study (p < 0.05; Table 2). Meanwhile, soil salinity could also influence the balance of phosphorus (Liu et al. 2012). The accumulating salinity in low-salinity soils will promote adsorption of AP to soil (Spiteri et al. 2008), which is supported by the results of this study (Tables 1, 2). Jun et al. (2013) also proved that the addition of salinity will promote P adsorption in freshwater floodplain forest soils of Georgia. Therefore, the seawater intrusion caused by global warming will contribute to the AP adsorption. The significantly positive correlation between soil phosphorus, nitrogen, and carbon indicated the intrinsic relationships between these three nutrients as they are the essential elements of organic matter and exist in organic matter in certain proportion (Wang et al. 2009). Aluminum, which formed in the course of soil development, was a strong determinant of TP and AP in soil as Al–P might be the important P fraction in the study area, which is consistent with the result of Satti et al. (2007), who presented that high P retention in soil is attributed to active Al associated with organic and mineral fractions (Igwe et al. 2010). Wang et al. (2012) also demonstrated that the Al oxyhydroxides have a small bulk and big surface provide much of the P sorption, resulting in the small AP/TP ratio in the study area. The regression analysis also exhibited the linear relationship between Al and AP or TP in this study. The significantly negative correlation between AP and soil depth also suggested the shallower distribution of AP. Higher clay and silt contents contributed to higher levels of AP and TPS, while higher sand contents might lead to soil P loss through leaching (Xiao et al. 2012). Therefore, soil properties such as SOM, salinity, silt, TN and Al were the main factors influencing soil AP and TP distribution.

Conclusions

We investigated spatial and seasonal distribution patterns of soil P content and stock and their influencing factors in the seasonal flooding wetland of the Yellow River Delta in two seasons in this study. TP levels in top soils in spring were slightly lower than those in fall, while AP levels in top soils in spring were obviously higher than those in fall. Profile distributions of soil P did not show significantly variability, but AP mainly accumulated in upper soils (0–20 cm) in this wetland. Compared to spring, the mean TP stock for each soil layer in this region was significantly higher in fall. Some soil properties such as SOM, salinity and silt with higher levels in soil could promote the increase in soil AP, and higher contents Al could enhance the accumulation of soil AP and TP. Therefore, the findings of this study indicate that soil AP in seasonal flooding wetlands can be controlled by regulating water and sediment in this region. However, the influencing mechanisms of these factors on soil P distributions and P fractions in soil profiles need to be further studied.