Introduction

Submerged macrophytes are a vital part of shallow lake ecosystems. They occupy the entire water column, from the bottom to the surface of a shallow lake. These plants convert sunlight into energy and make chemical elements bioavailable, allowing them to be absorbed by other plants (Qin 2009; Yu et al. 2010). They play an important role in phosphorus (P) cycling, especially in shallow lakes. They are very active recyclers of sediment P and should be viewed as potential P pumps (Carignan and Kalff 1980). They also have many other important ecological functions, such as improving dissolved oxygen (DO) content in the environment and changing the oxidation reduction potential (ORP) and pH of the water and sediments through photosynthesis (Moss 1990). They are primarily responsible for the maintenance of clear water in shallow lakes, and they often affect the functioning of the entire lake’s ecosystem (Scheffer and Jeppesen 2007).

During the growth period of submerged macrophytes, they can accumulate nutrients from both water and sediments. However, when submerged macrophytes die and decompose, photosynthetic production of oxygen ceases and organic matter and nutrients are released back into the aquatic environment. Bacteria and fungi that decompose decaying plant material in turn consume DO for respiration. Therefore, the control and management of submerged macrophytes are very important, especially after they have been restored in a lake (Li et al. 2014). The decomposition of aquatic macrophytes can substantially influence the long-term recycling of nutrients in freshwater ecosystems (Carpenter and Lodge 1986; Webster and Benfield 2003).

P present in water can be categorized into particulate phosphorus (PP) and dissolved total phosphorus (DTP). DTP can be subdivided into dissolved organic phosphorus (DOP) and soluble reactive phosphorus (SRP) (Jarvi et al. 2002). SRP is the most abundant form and the one most widely measured in natural waters, and it can be used by plants directly. DOP can play an important role in biological and biogeochemical processes. Previous studies have reported that phytoplankton are capable of assimilating P from the organic fraction, and DOP is reported to contribute a significant fraction of the DTP in natural waters, particularly where there is biological activity (Yoshimura et al. 2007). PP plays an important role in the change between different P fractions. In addition to containing orthophosphate, which can be directly absorbed and used by submerged macrophytes, it can transform into a type of dissolved P or be absorbed by the sediments (Gao et al. 2009).

The decomposition of submerged macrophytes is a complex process that is mediated by microorganisms. During the early decay phase of macrophytes, the leaching of water-soluble substances plays a key role in the loss of litter mass (Varga 2003). The rapid release of phosphorus accompanies this leaching. The amount of dissolved phosphorus released from plant cells into the water column increases rapidly during the first few days. P concentrations in the water reach their maximum value after about 15 days and subsequently decrease gradually over time (Carvalho et al. 2015; Chimney and Pietro 2006). During plant decomposition, hardly any gaseous phosphorus is released (Boulton and Boon 1991). Therefore, the released P must sink to the sediment at the bottom. During the decomposition phase of macrophytes, sediments and plant detritus absorb P simultaneously at an unknown rate. Little is known about the role that plant detritus plays in the stabilization of P during the decomposition of macrophytes in shallow lakes. Although it is known that sediments (containing detritus) play a predominant role in the decrease of P during the end of the decay phase of macrophytes (Juston et al. 2012), there is little information about the importance of detritus in P cycling during the decomposition of macrophytes in shallow lakes.

There have been numerous recent studies of P release during the decay phase of submerged macrophytes in natural and constructed wetlands (Asaeda et al. 2000; Titus and Pagano 2002; Wrubleski et al. 1997). Most of these studies use the litter bag technique to measure the rate of macrophyte decomposition. Using this technique, macrophytes are collected from the study area and dried at 60 °C or air-dried to constant mass and placed in plastic bags. If the macrophytes are large plants, such as Phragmites australis, they are usually fragmented to a coarse litter containing natural proportions of stems and leaves (Ágoston-Szabó et al. 2006; Longhi et al. 2008). Decomposition rates are estimated with an exponential model (Olson 1963):

$$ W_{t} = W_{0} {\text{e}}^{ - kt}, $$
(1)

where t is the time (in days) since the start of the experiment, Wt is the dry mass of litter remaining at time t (%), W0 is the initial dry mass of litter at time 0 (defined as 100%), and k is the decomposition rate coefficient (day−1).

We determined that the litter bag technique cannot effectively estimate the rate of macrophyte decomposition. Only part of the plant cell dies during decomposition, and generally the decomposition progresses from the leaves to the stem. The litter bag technique assumes that all plant cells die at the beginning of the decomposition phase (Smock and Stoneburner 1980), so the decomposition rate that it generates may not be accurate.

In the present study, we evaluated the decomposition process for Potamogeton crispus in a microcosm. The aim of the study was to determine the decomposition rate of P. crispus during natural senescence and to gain insights into the cycling of P between water and P. crispus.

Materials and methods

Study area

Aboveground parts of the submerged macrophyte, P. crispus, were collected from Lake Yimeng in Linyi in China (35°05′N, 118°19′E) at the beginning of the decomposition season. Lake Yimeng was formed in 1997 when a rubber dam (1135 m) was built across the Yi River, capturing about 17 km2 of water. In recent decades, the lake has exhibited dense canopy-forming populations of P. crispus, which covers nearly 90% of the lake during the spring and summer. Different types of management have been applied to reduce algal blooms (eutrophication), including harvesting P. crispus in summer and applying algaecides. However, in recent years, the lake has exhibited severe algal blooms in both summer and early autumn (Fig. 1).

Fig. 1
figure 1

Lake Yimeng: Potamogeton crispus in spring; algal blooms after the death of P. crispus in summer

Full size image

The amount of P loading from the surrounding watershed was 90.31 t each year (Linyi Hydrological Bureau 2016; Wang et al. 2007). The average biomass of P. crispus was 7.98 kg m−2; therefore, the estimated maximum amount of P stored in P. crispus was 39.07 t (2.55 g m−2) in this lake. P. crispus stored 43.26% of the P in the lake. Almost all of the P was released back into the water during the decomposition period, and only a few aquatic plants (e.g., Nymphoides peltata and Trapa bispinosa) grew in this period, covering about 5% of the lake area. The decomposition of P. crispus may be one of the important triggers for algal blooms in this lake. The average TP concentration was 0.13 mg L−1 from March to June, which is the growth period of P. crispus, and it was 0.56 mg L−1 from July to September, which is the decomposition period of P. crispus.

Experimental microcosm

In order to determine the rate of P released by P. crispus during decomposition, plants were collected from different sites in Lake Yimeng on May 15–16. Plant materials were partially living at the beginning of the experiment. As the vegetation was heterogeneously distributed in the lake, subsamples were randomly taken from different parts of the lake. The samples covered the center (depth 1.5–2.3 m), the shore (depth 1–1.3 m), and the southern and northern parts. In each sample site, the sampling area was delimited using a polyvinyl chloride tube (area 0.50 m2), and completely submerged macrophytes were collected from the boat. A rake was used to remove plants from the bottom, collecting both aboveground and belowground biomass (AGB and BGB, respectively). In order to determine the ratio of BGB to AGB, the weights of BGB (root and rhizome) and AGB (stem, turion, and leaf) of P. crispus were recorded for each plant. The average ratio of BGB to AGB was 0.059 ± 0.04 (dry weight). Because BGB made a very low contribution to total biomass, and as it was difficult to collect all the parts of P. crispus from the lake bottom, only the AGB of P. crispus was used in the microcosm experiments. Most of the AGB of P. crispus breaks off from the root when it begins to decompose at the end of the growing season. Therefore, in the present experiment, only the aboveground portion was used to simulate the course of decomposition.

Harvested P. crispus was incubated with lake water in a tank for less than 1 day before the adherent materials were removed. Then the plant materials were gently rinsed with tap water and distilled water to remove sediment and periphyton. P. crispus was divided into 99 parts, each with a fresh weight of 100 g. Each fresh sample of P. crispus was placed in a beaker (4 L), to which was added 3.5 L deionized water. The system simulated the decomposition course of P. crispus in still water with no circulation. The decomposition experiment proceeded in an incubator with a stable temperature of 25 °C (± 3 °C). The light intensity was 800 ± 50 µmol m−2 s−1 (photoperiod 12:12 h light:darkness). The experiment spanned 30 days. Each day, three breakers were collected and the biomass (fresh weight and dry weight) and P concentrations in water and plant material were recorded. The values of fresh weight, dry weight, and P concentration in water and plants reported here are the means of triplicate samples. The remaining mass (%) is based on the dry weight.

Water and plant analyses

The pH (PHSJ-4A, Lei-ci, Shanghai, China), ORP (Lei-ci), and DO in the overlying water (5750, YSI, Yellow Springs, OH, USA) were measured first.

Next, 10-mL water samples were removed from the three beakers that were collected daily and stored in a refrigerator at 4 °C until TP, PP, DTP, DOP, and SRP analysis. The methods are summarized below. For TP analysis, the water sample was autoclaved at 121 °C for 30 min after K2S2O8 was added. Thereafter, 10% ascorbic acid was added, and the sample was measured using the molybdenum blue spectrophotometric method. A continuous flow analyzer (Flowsys III, Systea Company, Anagni, Italy) was used to determine the P concentration. The same method was used for the SRP determination, except that the water sample was filtered through a 0.45-μm cellulose acetate membrane and not autoclaved. DTP was measured using the same method as that employed for TP, but again the water was filtered through a 0.45-μm cellulose acetate membrane. The difference between TP and DTP was defined as the sum of the PP fraction. The difference between DTP and SRP was defined as the sum of the DOP fraction. The detection limit for SRP and DTP concentrations in the overlying water was 1 μg L−1. All materials used for these analyses were purchased from Shanghai N&D Co. Ltd. (Shanghai, China). For all samples, triplicates were analyzed, and the data are expressed as the mean.

An appropriate amount of deionized water was then added to other beakers to compensate for the loss of water through evaporation every 10 days. The following equation was used to calculate the P concentration:

$$ c_{n} \; = \;\frac{{c_{n}^{\prime } v}}{v\, + \,v_n}, $$
(2)

where Cn is the actual P concentration at the n-th sampling time, \( c_{n}^{\prime } \) is the measured P concentration before adding deionized water, v is the volume of overlying water before adding deionized water, and vn is the volume of deionized water added at time n.

Water samples were siphoned from approximately 10 cm below the surface of the beakers (i.e., around the middle of the beakers). The samples were kept at 0–4 °C and P was analyzed within 24 h of sampling.

P. crispus from each beaker was oven-dried at 85–90 °C and maintained in an air-circulating oven for approximately 2 weeks until it achieved a constant dry weight (DW). The dry material was ground and stored in vacuum-sealed bags at room temperature. The concentrations of P, nitrogen (N), and carbon (C) in plant material were determined by standard methods (Sommers 1977).

Data analysis

The results are presented as averages of triplicates. The significance of treatment effects was determined at the 0.05 probability level. SPSS 16.0 and SigmaPlot 12.5 were used for model analyses. The mass loss coefficient (k) was calculated for a whole period, based on the exponential decay model (Eq. 1).

Results

Mass loss and P concentration in plants

The mass loss coefficient k, based on the exponential decay model (Eq. 1), was 0.05 day−1. P. crispus retained 22.00% (all dead material) of its original weight after 30 days of incubation. The mass changed rapidly in the first 10 days (declining to 50.00% of the original mass, with partially living material) and changed slowly in the last 20 days (from 50.00 to 22.00% of the original mass).

Therefore, the first 10 days appear to be the rapid decomposition phase, whereas the last 20 days represent the slow decomposition phase (Fig. 2). Variation in the P concentration in P. crispus was observed during decomposition (Fig. 2). The P concentration in P. crispus detritus increased slowly during the first 22 days (from 3.90 to 3.98 g kg−1 DW), but accelerated starting on day 23. After 30 d of incubation, the P concentration in P. crispus detritus was 5.20 g kg−1.

Fig. 2
figure 2

Phosphorus (P) concentration (mean ± SD) in plants and the mean (± SD) percentage of Potamogeton crispus mass remaining after decomposition during the study period

Full size image

P concentration in water

The changes in P concentration in water during decomposition are shown in Fig. 3. TP, DTP, and SRP increased quickly during the first 15 days and decreased slowly from day 16. On the 10th day of the decomposition experiment, the concentrations of TP, DTP, and SRP were 6.32, 5.60, and 4.48 mg L−1, respectively. By the end of the experiment, TP, DTP, and SRP reached 5.57, 5.10, and 4.13 mg L−1, respectively. Therefore, the first 10 days constitute a period of rapid P release, whereas days 10–15 represent a period of slow P release. Compared with TP, the DTP and SRP concentrations of DOP and PP remained at low levels (DOP 0–1.12 ± 0.06 mg L−1, PP 0–1.81 ± 0.02 mg L−1). DOP and PP increased slowly during the decomposition phase, after which DOP showed a slight decrease, starting on day 15. The maximum values of the different P fractions in water could be ordered as follows: DTP > SRP > PP > DOP.

Fig. 3
figure 3

The phosphorus (P) concentrations (mean ± SD) in water, including total P (TP), dissolved total P (DTP), particulate phosphorus (PP), soluble reactive phosphorus (SRP), and dissolved organic phosphorus (DOP)

Full size image

In this experiment, P release occurred from the beginning to the maximum value of P, and the P sedimentation phase occurred from the maximum value of P to the end of the experiment. The P release rates of TP, DTP, SRP, DOP, and PP were 0.1389 ± 0.0021, 0.1307 ± 0.0020, 0.1045 ± 0.0011, 0.0261 ± 0.0008, and 0.0253 ± 0.0015 mg day−1 g−1, respectively (based on the dry detritus weight recorded at the beginning of the experiment). The P sedimentation rates of TP, DTP, SRP, DOP, and PP were 0.0641 ± 0.0031, 0.0700 ± 0.0028, 0.0597 ± 0.0013, 0.0081 ± 0.0010, and 0.0013 ± 0.0001 mg day−1 g−1, respectively (based on the dry detritus weight recorded at the beginning of the experiment).

The various phosphorus fractions and environmental factors were correlated (Table 1). TP, DTP, SRP, and DOP were negatively correlated with the environmental factors pH, DO, and ORP. PP was negatively correlated with pH and DO but weakly positively correlated with ORP, TP, DTP, SRP, and DOP. In addition, TP, DTP, SRP, DOP, and PP were negatively correlated with the amount of mass remaining.

Table 1 Pearson correlation coefficients between various phosphorus fractions and environmental factors
Full size table

The nitrogen concentration in P. crispus decreased quickly during the first few days (from the 1st to the 9th day) and gradually increased at the end of the experiment (to 31.53 ± 0.63 g kg−1 on the 31st day). The carbon concentration in P. crispus remained stable from the 1st day to the 20th day, but decreased from day 21, to reach 370.00 ± 18.5 g kg−1 at the end of the experiment (Fig. 4).

Fig. 4
figure 4

The nitrogen and carbon concentrations (mean ± SD) in Potamogeton crispus during the study period

Full size image

All P concentrations in water increased to the maximum value and then decreased gradually with time. A peak equation (Wang 2015) was used to describe the changes in P concentration in water:

$$ P_{t} = \frac{a}{t}{\text{e}}^{{\left[ { - 0.5\left( {\frac{\ln (t/k)}{b}} \right)^{2} } \right]}}, $$
(3)

where t is the time (in days) since the start of the experiment, Pt is the P concentration in water at time t (mg L−1), and a, k, and b are constants.

Values of the regression coefficient of determination for and the constants in Eq. 3 are shown in Table 2. TP, DTP, SRP, DOP, and PP had high values of R2. These high R2 values suggest that they can be used to predict a model of P release during P. crispus decomposition.

Table 2 Values of the regression coefficient of determination for and the constants in Eq. 2
Full size table

Environmental factors in the experiment

The environmental factors in the water changed over time as decomposition progressed. ORP and DO decreased quickly during the first few days and gradually increased at the end of the experiment, whereas pH gradually decreased during the course of the experiment (Fig. 5). The temperature remained stable at approximately 25 °C. The pH gradually decreased from neutral (7.10 ± 0.35) at the beginning of the experiment to slightly acidic values (6.10 ± 0.31) at the end. ORP decreased quickly during the first few days, reaching − 110.0 ± 2.2 mV on day 9. ORP increased gradually to 22.5 ± 1.3 mV on day 25 and then remained stable, followed by a slight decrease to 12.1 ± 0.6 mV by the end of the experiment. DO decreased quickly over the first few days, reaching 1.01 ± 0.21 mg L−1 on day 13. After that it remained stable, and then it increased from day 24, reaching 5.02 ± 0.25 mg L−1 by the end of the experiment.

Fig. 5
figure 5

The temperature (T), pH, oxidation reduction potential (ORP), and dissolved oxygen (DO) (mean ± SD) in water during the study period

Full size image

Discussion

The final decomposition rates measured for P. crispus differed from those reported in other studies that examined the decomposition of dried plants, such as k = 0.0205 day−1 in Howard-Williams and Davies (1979). Our results were closest to those reported by Rogers and Breen (1982); i.e., k = 0.04 day−1 for natural senescence and k = 0.02 day−1 for dry materials. Moreover, the decomposition rates of P. crispus from the present study differed from those of other species of Potamogeton, including decomposition rates of 0.01 day−1 reported for P. maackianus (Xie et al. 2004) and 0.019 day−1 for P. pectinatus (Carvalho et al. 2015). The decay rate for each species showing uniform exponential decay can be calculated as a constant for the whole decay period from Eq. 1. The results of this comparison are shown in Table 3. The overall rate of mass loss was slower from dried P. crispus than from senescent plants.

Table 3 Comparison of the decomposition rates of various species of aquatic macrophytes
Full size table

P. crispus is distributed in a wide range of climatic regions. It can tolerate hypertrophic conditions and grows well in polluted water (Heuschele and Gleason 2014). P. crispus plays a central role in aquatic ecosystems by helping in the regulation of nutrient levels, including P. The growth of P. crispus can lead to P and N depletion and decreased chemical oxygen demand, along with increased water transparency and DO in the water column (Bakker et al. 2010). These characteristics make it a suitable candidate for bioremediation of polluted waters, especially in shallow lakes (Leoni et al. 2015). Previous studies indicated that P. crispus has the capacity to regulate nutrients in shallow lakes by absorbing large amounts of nutrients.

Our results showed that P. crispus had a high degradation rate. Field monitoring of Lake Yimeng showed that the TP concentration increased from April to June, reaching an average concentration of 0.27 ± 0.02 mg L−1 in June (Fig. 6). The highest TP concentration was lower than that in our microcosm experiment. P absorption by sediments, algae utilization of P, and harvesting of P. crispus may be the main underlying reasons for this. The TP concentration of Lake Yimeng decreased in July because water levels were high during the rainy season. The high rates of degradation found in our experiment indicate that this plant plays an important role in nutrient dynamics in the lake studied.

Fig. 6
figure 6

The TP concentrations (mean ± SD) in Lake Yimeng over the course of a year (n = 12)

Full size image

Several factors affect the rate of degradation of aquatic macrophytes in lentic environments, including the chemical composition of the plant and the activity of the microbial decomposer community (Gessner 2000). Detritus with high N and P concentrations and a low C concentration has low C:N and C:P ratios and is considered good-quality detritus. König et al. (2013) studied leaf breakdown for different macrophytes in a subtropical stream and reported that those with a higher nutritional content might be more palatable to scavengers, which would positively influence the rate of tissue degradation.

Our results showed that the N and P contents increased and the C content decreased by the end of the decomposition phase (Figs. 2, 5). The increase in the P content is generally attributed to uptake by decomposer microbes associated with the plant tissue and net P immobilization by the decomposing plant tissue (Gonçalves et al. 2004). At the end of the experiment, the plant detritus could be categorized into a larger plant detritus part and a flocculation part. Only the larger detritus part was analyzed. The flocculation part included sedimentary P and other P fractions; this part of the P was equivalent to the whole P minus the P in water and the larger plant detritus (Fig. 7). The percentages of P in water, the larger plant detritus, and the flocculation part corresponded to 32.30, 29.34, and 38.36% of the total, respectively, at the end of the experiment. These results are similar to those of other studies of macrophyte decomposition in lakes (Byren and Davies 1986; Chimney and Pietro 2006; Wrubleski et al. 1997).

Fig. 7
figure 7

The percentages of phosphorus (P) in water, plants, and sediments and other fractions

Full size image

The increased N and P concentrations and decreased C concentration in the detritus led to low C:N and C:P ratios by the end of the decomposition phase. Menendez et al. (2004) reported that plant material with a higher nutritional content might be more palatable to scavengers. C:N and C:P ratios are considered to be two important factors affecting microbial activity and population dynamics, in particular by inhibiting the immobilization of N and P by microorganisms (Dahroug et al. 2016). Studies of submerged macrophytes in lakes have shown that the decomposition rate is mainly associated with the bacterial community (Millelindblom and Tranvik 2003). The reductions in C:N, C:P, and N:P (Fig. 8) in the present study revealed an increase in nutritional quality during decomposition. The regulation of nutrient balance by P. crispus depends on its life stage (growth vs. senescence). P. crispus plays a double role with respect to the P cycle and may act as a nutrient sink (Siong and Asaeda 2006).

Fig. 8
figure 8

The carbon:nitrogen (C:N), carbon:phosphorus (C:P), and nitrogen:phosphorus (N:P) ratios, based on mean values of C, N, and P in Potamogeton crispus plants during the study period

Full size image

The relatively high water temperature (25 ± 1 °C) probably accelerated the decomposition of the vegetation, causing decreases in the DO, pH, and ORP of water (Fig. 4) that coincided with P release (Asaeda et al. 2000).

The sharp increase in the P content of P. crispus during the first 10 days of decomposition suggests higher internal P loading via release through submerged macrophytes (Kröger et al. 2007). The majority of the P released from plants (approximately 80%) occurs in soluble form. Thus, the major P pathway is from the sediments to the water via macrophytes. This hypothesis was confirmed by our results, which showed that the physical leaching of P from plant tissue occurred during the early stages of senescence.

Our results indicated that the P concentrations in the overlying water mainly depend on the plant P content and developmental stage. P. crispus removes P from the water and sediment during its growth period; however, the rapid decomposition rates suggest that much of the accumulated nutrients will eventually be returned to the water ecosystem. Therefore, the harvesting of P. crispus prior to its senescence may significantly reduce the internal P load (Bennett et al. 2001). Furthermore, the management of P. crispus by other means, such as the application of algaecides, could reduce the occurrence and intensity of algal blooms (Boström et al. 1988).

Conclusion

We measured decomposition rates of and P release from the submerged macrophyte P. crispus in a microcosm experiment. Using the exponential decay model, the P. crispus mass loss coefficient k was found to be 0.05 day−1 under conditions of natural senescence. All P fractions in water reached a maximum value and then decreased gradually. The maximum values of the different fractions in water could be ordered as follows: DTP > SRP > PP > DOP.

P was released from the beginning of the present experiment until it reached its maximum value. Thereafter, P was absorbed until the end of the experiment. The P release rate of TP was 0.1389 ± 0.0021 mg day−1 g−1, and the P sedimentation rate of TP was 0.0641 ± 0.0031 mg day−1 g−1.

The environmental factors pH, DO, and ORP declined along with the rapid P release. The P and N contents in water had increased by the end of the experiment, whereas the C:N, C:P, and N:P ratios had decreased. The high N and P concentrations and the low C concentration of P. crispus are the main reason for the rapid decomposition.

Our findings provide insights into the course of decomposition and indicate that effective management strategies for controlling eutrophication in lakes that harbor P. crispus could include the harvesting of this submerged macrophyte before its decomposition.