Inorganic phosphorus is one of the critical nutrients determining trophic state and freshwater productivity. Sediment may act as a sink or source of phosphorus to the overlying water depending on its pH, redox state, various forms of phosphorus present, etc. To examine potential sorption or mobilization of sediment phosphorus in floodplain wetlands, the amount and distribution of phosphorus fractions were evaluated using a sequential chemical extraction procedure. Exceedingly high levels of total phosphorus (mean: 6040 ± 344, 5470 ± 363 mg kg−1), consisting largely of organic and refractory fraction (70 – 98%), followed by calcium-phosphorus (mean: 584 ± 31.3, 143 ± 8.42 mg kg−1) and iron-phosphorus (mean: 108 ± 10.1, 91.0 ± 7.68 mg kg−1) were recorded respectively in Bhomra and Akaipur wetlands of West Bengal, India. The inorganic phosphorus, comprising the loosely sorbed phosphorus and all the mineral bound forms contributed only about 6–14% to the total phosphorus indicating their less significance in phosphorus sorption or desorption in these tropical wetlands. Although the loosely sorbed phosphorus was in moderate level (2.69 ± 0.69, 1.54 ± 0.53 mg kg−1), water dissolved phosphorus was recorded at higher concentrations (mean: 0.16 ± 0.02 mg l−1 in Akaipur and 1.08 ± 0.12 mg l−1 in Bhomra). However, the higher level of water available phosphorus was not reflected in plankton production since the dominance of weeds suppresses their growth. This study recorded large accumulation of organic matter and nutrients in the form of detritus in these wetlands which may be channelized for fish production through stocking of suitable detritivorous fishes and/or reducing macrophyte coverage that would give space and nutrients for phytoplankton growth.
Phosphorus (P) is one of the limiting nutrients for phytoplankton growth in some freshwater ecosystems (Boström et al., 1982). Being negatively charged, phosphate tends to bind with soil cations like Ca, Mn, Fe, and Al. The bound forms are poorly soluble (Bagyaraj et al., 2000) and only a small fraction of sediment P remains biologically available. A large portion of the P naturally present, as well as part of the fertilizer P applied in aquaculture or that coming through run-off, etc. gets immobilized in sediment and become unavailable to primary producers (Jana, 2007). Besides P sorption, a significant amount of P is also released from sediment to the overlying water, especially during summer, triggering seasonal eutrophication (Boström et al., 1988; Søndergaard et al., 1999). The ability of sediment to act as a sink or a source of phosphorus depends on the physical, chemical, and biological characteristics of the water body. In presence of high level of Ca and alkaline pH, sediment can effectively bind P (Gonsiorczyk et al., 1998). In acidic environments, especially at the sediment pH below 5.5, both Fe+3 and Al+3 precipitate phosphorus (Banerjea and Ghosh, 1970). While iron (Fe+3) binds P under oxic-condition, Fe+2 has less affinity and thereby releases iron-bound P in a reduced environment (Boström et al., 1988). Aluminum bound P is stable and less mobile than iron bound P (Ann et al., 2000). Phosphorus fractionation gives valuable information about potential internal loading or fixation of P in sediment and is a useful tool in long-term trophic state management of lakes and rivers (Psenner et al., 1984; Renjith et al., 2011).
Floodplain wetlands, particularly ox-bow lakes, are integral parts of riverine ecosystems. These ecosystems provide substantial fish to local communities, and play vital ecological functions like water reserve and recharge, flood control, purification, etc. India has 2.02×103 km2 of floodplain wetland area, concentrated mostly in the states of Assam, West Bengal, Bihar and Uttar Pradesh (Sinha, 2003). The floodplain wetlands of West Bengal have neutral to alkaline sediments in comparison to the predominantly acidic sediments in the wetlands of Assam due to water shed characters and climatic variability. These resources are rich in endemic fish species and considered highly productive at 1000–1500 kg fish ha−1. However, actual production from these wetlands is often <200 kg fish ha−1, except in some well managed wetlands where production exceed 800 kg fish ha−1 (Vinci, 2003). The low productivity is mainly due to improper management, weed infestations and nutrient imbalances, mostly that of P and N (Das, 2003). To address the nutrient related issues and poor productivity of the floodplain wetlands, accounting of P availability is very important. The objective of this study was to assess various P fractions in floodplain wetland sediments to have an estimate of the nutrient composition and availability in these water bodies. The data generated will also help in explaining the observed trophic state and managing the ecosystems in a sustainable manner maintaining optimum production level.
Materials and methods
The present study was conducted during March 2009 to February 2012 in two floodplain wetlands, namely ‘Akaipur’ (23°04′N, 88°43′E) in Nadia district and ‘Bhomra’ (22°59′N, 88°38′E) in North 24 Parganas district of West Bengal, India. They are closed ox-bow lakes in the Gangetic basin representing wetlands having acidic and alkaline sediments respectively (Figure S1 in the online supplementary information). Stocking and regulated capture fishery is practiced in both the wetlands providing about 600–1000 kg ha−1 fish yield and livelihood support to the fisher community year round (Sugunan et al., 2000). While the Water Hyacinth, Eichhornia crassipes, is a major problem in Akaipur, submerged macrophytes cover a major part in Bhomra wetland. The species include Hydrilla verticillata, Ceratophyllum demersum and Najas sp. Both wetlands receive allochthonus inputs, especially during rainy season from surrounding agricultural fields, jute retting, animal husbandry and domestic effluents, etc., but there is no industrial or city waste inflow (Sugunan et al., 2000).
The top 0–14 cm layer of bottom sediments were collected using Ekman dredge from five sampling stations (Figure S1), covering both littoral and profundal zones, in each wetland. Water samples were collected using Niskin water sampler from 0.5 m water depth in sterile dark bottles properly cleaned with P-free detergent. The water and sediment samplers were thoroughly washed with wetland water from the sampling area to avoid cross contaminations. All the samples were transported under ice cover and processed within 3–4 h of sampling.
Estimation of physico-chemical parameters of water and sediment
The pH and specific conductance of water and sediment were measured using multiparameter water analysis probe (Multiline P4, WTW, Germany); sediment samples were air dried, ground, sieved (10 mesh, 2 mm) and suspended in Milli Q grade water at the ratio of 1:2.5 before pH and conductivity measurements. Available P concentration in water was measured spectrophotometrically using a commercial Phosphate Test Kit (Spectroquant® Phosphate Test, Merck KgaA, 64271 Darmstadt, Germany). Total organic carbon (TOC) of sediment was determined following the method of Walkley and Black (1934).
Sediment P fractionations
Different P forms in ground and sieved sediment samples were estimated by sequential fractionation following the methods of Jackson (1973). Briefly, successive extraction of initial 2 g sieved sediment sample, in triplicate, was done with 1N NH4Cl, 0.1 M neutral NH4F and 0.1 M NaOH to extract loosely sorbed P (Avail-P), Al-P and Fe-bound P respectively, and finally, 0.1 M H2SO4 was used to extract Ca-bound P. For extraction of each fraction, samples were shaken in a horizontal shaker for 2 h, kept overnight, then the supernatant was decanted and the residue was treated for the next fraction. The reactive P in the supernatant was measured by the molybdate-blue method of Murphy and Riley (1962). After Al-P extraction, sediment residues were washed twice with saturated NaCl solution for 2 min before extraction of the next fractions. For total-P estimation, 0.5 g of dry sample was sequentially digested with 40 ml of 30% H2O2 and tri-acid mixtures (nitric acid, perchloric acid, sulfuric acid in ratio of 10:4:1 V/V) for two consecutive days, solubilized with water and P concentration estimated. The residual-P, consisting of organic and refractory P, was calculated by subtracting the total inorganic P from the total P content. In each sampling, the mean of five sampling sites in a wetland was taken as one datum.
Variations in loosely sorbed P, Fe-P, Al-P, Ca-P levels from the wetlands were examined by ANOVA and subsequent post-hoc analysis was performed using SAS Enterprise Guide (4.2). Correlation between sediment exchangeable P (Avail-P) with water available-P, TOC and residual P were examined by bivariate fit using JMP® Genomics 4.1 of SAS. The results include observed range, arithmetic mean ± standard error.
Water and sediment characteristics of the wetlands
The water samples of Akaipur wetland were neutral to mildly acidic and the sediment samples were moderately acidic (Table 1). Both water and sediment samples from Bhomra were mild to moderately alkaline (Table 1). Available P ranged from 0.25–1.81 mg l−1 (mean: 1.08 ± 0.12 mg l−1) in Bhomra and 0.03–0.30 mg l−1 in Akaipur (mean: 0.16 ± 0.02 mg l−1). Conductivity of water and sediment were also higher in Bhomra. Sediments were rich in organic matter with higher TOC content in Bhomra (Table 1).
Sediment phosphorus content and forms
Wetland sediment samples were rich in total P with ranges of 1950–8410 mg kg−1 (mean: 5470 ± 363 mg kg−1, n 36) in Akaipur and 3200–8530 mg kg−1 (mean: 6040 ± 344 mg kg−1, n 36) in Bhomra (Figure 1a). However, levels of exchangeable P were low to moderate, ranging from 0.04–4.12 mg kg−1 in Akaipur and 0.04–6.21 mg kg−1 in Bhomra (Figure 1b). The levels of exchangeable P only accounted for 0.03–0.04% of total sediment P load and were recorded to be weakly correlated with the water available P (r2 = 0.27, p < 0.005). The Ca-P ranged between 31.1–198 mg kg−1 (mean: 143 ± 8.42 mg kg−1) and 385–946 mg kg−1 (mean: 584 ± 31.2 mg kg−1), constituting only 0.41–7.54% and 3.55–19.6% of total sediment P in Akaipur and Bhomra, respectively. The Ca-P level was significantly higher in Bhomra (p < 0.01).
The NaOH-P, representing the redox-sensitive P fraction bound to Fe-hydroxides and Mn compounds (Kozerski and Kleeberg, 1998), ranged from 16.2–151 mg kg−1 (mean: 91.0 ± 7.68 mg kg−1) in Akaipur and 20.2–187 mg kg−1 (mean: 108 ± 10.1 mg kg−1) in Bhomra. Unlike Ca-P, the Fe-P content was more or less the same in both the wetlands, contributing only about 1.66–1.78% to the total sediment P load. The Al-P level ranged widely between 29.9 –190 mg kg−1 (mean: 96.9 ± 13.4 mg kg−1) in Akaipur and 10.4 –447 mg kg−1 (mean: 128 ± 27.0 mg kg−1) in Bhomra (Figure 1b), with insignificant differences between the wetlands. The Al-P shared 1.77–2.13% of total sediment P. Total inorganic-P, defined as the sum of Avail-P, Al-P, Fe-P and Ca-P, contributed only 6.07% and 13.6% to total P in Akaipur and Bhomra, respectively (Figure 1a). The Residual-P, composed of organic and refractory P, contributed 69.9–97.8% (average: 89.5 and 83.8% in Akaipur and Bhomra, respectively) of the total sedimentary P. The mean organic and refractory P content was 4890 ± 332 mg kg−1 in Akaipur and 5060 ± 303 mg kg−1 in Bhomra (Figure 1a). Difference between the two wetlands was statistically insignificant.
Among the inorganic fractions, the Ca-P was most dominant (about 71%) in Bhomra, while Fe-P and Al-P jointly contributed about 50% and Ca-P fraction was about 48% in Akaipur (Figures 1c and d). The order of different P fractions were: Residual P (refractory and organic) > Ca-P > Fe-P > Al-P > Avail-P in Akaipur and Residual-P (refractory and org-P) > Ca-P > Al-P > Fe-P > Avail-P in Bhomra wetland.
The studied wetlands sediments were very rich in total P content. In Akaipur, they were as high as 9490 mg kg−1 and up to 11,700 mg kg−1 in the rainy season (selected from the full data set pre-averaging). These were much higher than those reported by others (Banerjea and Ghosh, 1970; Rzepecki, 2010; Renjith et al., 2011). The average sediment C:P ratios were also low (C:P = 5.39, 6.79 by mass for Akaipur and Bhomra, respectively) indicating P enrichment. It is estimated that nearly 50% of the P inputs from fertilizer, organic manures, etc. is retained in agricultural soil after crop harvest (Pathak et al., 2010) and a significant portion of it was transported to aquatic environments through storm water. In absence of riverine flow and sewage inputs, we presume that the contribution of P from soils in the catchment area with burgeoning rice and vegetable cultivation, contributes substantially to the total P burden in these wetlands, with minor inputs from households, animal excreta, etc. Although the loosely sorbed P was an insignificant fraction of total P in sediments, the levels of water soluble reactive phosphorus (SRP) were moderate to high in terms of tropic status (Carlson and Simpson, 1996). Taking into account the high levels of SRP, total P in sediment, water and sediment conductivity, nitrogen (nitrate) and types of macrophytes dominating the wetlands, we classify these wetlands as eutrophic (Carlson and Simpson, 1996; Das, 2003; Rolon and Maltchik, 2006; Feijoóa and Lombardob, 2007). Bhomra was more nutrient enriched which could lead to higher fish production. However, fish production in Bhomra was only marginally higher due to reduced catch per unit effort in macrophyte covered bottom and other management factors, like the extent of fish stocking, etc. Comparison of the reported levels of phosphorus in this study with historical data (Sugunan et al., 2000) indicates that the phosphorus availability has increased by a factor of 10 in the last 15 years, which is a marked enrichment in recent times. The moderate to high level of labile P in the sediment was not, however, always reflected in P content in the water. Available P levels in the water were sometimes low or at trace level and the disparity between high sediment available P and low water available P in some occasions could be explained by heavy P uptake by massive macrophyte production in the wetlands. This also suppressed the phytoplankton growth (Mitra, 2003).
Bio-available P is the sum total of immediately available phosphate and P that can be mobilized from other forms. Iron bound P is considered as a potentially mobile pool with P release occurring through reductive dissolution of Fe (III) oxyhydroxide and co-precipitated P under anaerobic conditions (Gunnars et al., 2002). Much of internal P loading to the overlying water occurs from this form in the summer (Kozerski and Kleeberg, 1998). In Akaipur and Bhomra the Fe-P contents were low at only 1.66–1.78% of total P, indicating either low Fe content in these organic matter rich wetlands and/or depletion of Fe-P due to development of anoxic bottom environment due to high rate of microbial decomposition (Bostrom et al., 1988).
Phosphorus associated with Ca2+ phase was the main inorganic-P pool. Apatite formation is a pH dependent process with higher Ca-P levels found in alkaline environments (Fytianos and Kotzakioti, 2005; Padma and Nair, 2010). This probably explains the high Ca-P reserve in alkaline Bhomra sediment and low Ca-P reserve in acidic Akaipur sediment. Apatite, being less influenced by redox changes, is considered to be a somewhat stable P fraction (Gonsiorczyk et al., 1998; Ann et al., 2000; Manna et al., 2014), although, microbial dissolution by organic acids could be a potential mechanism of apatite mobilization (Maitra et al., 2015). The Ca-P fraction was the prime form governing P sorption and desorption in marine and estuarine sediments and less so in riverine systems (Khalil et al., 2007; Padma and Nair, 2010). Although the recorded Ca-P concentrations are similar to that in the southern Indian river (Padma and Nair, 2010) or in eutrophic water bodies in China (Qian et al., 2010), the levels were much lower than those recorded in rivers, estuaries, coasts and sea, indicating less significance of Ca-P in overall sorption or release in these low alkaline to acidic freshwater environments.
The Ca-P form generally predominates in alkaline pH, while Fe-P and Al-P forms dominate in acidic environments (Chang and Jackson, 1957; Banerjea and Ghosh, 1970). We also observed higher proportions of Fe-P and Al-P in the acidic Akaipur than in the more alkaline Bhomra, although absolute levels of these fractions were higher in Bhomra by virtue of higher total P content. Phosphorus bound to Al or in Al-organic complexes are stable, being insensitive to redox changes and microbial dissolution (Ann et al., 2000; Fankem et al., 2006). This fraction contributed 1.77% of total P in Akaipur and 2.13% of total P in Bhomra. Overall, total mineral bound P shared only 5.89–13.6% of total sediment P, indicating its less critical role in overall P dynamics in these wetlands. As observed in Bhomra, calcium compounds might determine the availability of inorganic P in an alkaline aquatic environment, whereas in an acidic environment like Akaipur, Al and Fe control the P solubility. However, surpassing these fine regulations, organic matter plays major roles in P fixation/dynamics in these organic matter rich wetlands.
Residual P, comprising of refractory and organic forms, formed as high as 93.9% of total P in Akaipur and 86.4% in Bhomra, indicating that much of the sediment P was buried as organic matter in these tropical wetlands. This was substantiated by the high level of sediment organic carbon content. The observed dominance of organic-P or residual-P was similar to other organic matter rich environments (Fytianos and Kotzakioti, 2005; Rzepecki, 2010). Possibly, P as a part of humic substances and inorganic P bound with humus-metal complexes (Rzepecki, 2010) yielded high residual P contents, as well as low C:P ratio in these wetlands. However, presence of moderate to high levels of labile P in top layers of sediment indicated significant P regeneration and exchange along the sediment-water interface and that the residual P fraction might not be totally refractory in nature. The TOC correlated well the residual P (r2 = 0.58, p < 0.01) in Akaipur, but not in Bhomra, possibly due to presence of more non-exchangeable humus bound-P in Akaipur. In absence of allochthonus matter loading due to closed nature of these wetlands, and low phytoplankton growth much of the sediment organic matter originates from growth of floating and submerged macrophytes, which cover as much as 65–85% of the wetland area in this Ganga-Brahmaputra river basin (Hassan, 2003). Das (1998) estimated that macrophytes, mostly submerged, generated a coarse detritus load of 190 g m−2, which is rich in available N (1170 mg N kg−1 detritus) and available P (29.5 mg P kg−1 detritus) in such wetlands. The Akaipur wetland had a large floating Eichhornia biomass, while the Bhomra was covered more intensively by submerged macrophytes. Water Hyacinth is poor in N, P and less degradable (Reddy and DeBusk, 1991; Hassan, 2003) and thus, contributing more refractory organic matter to the sediment in Akaipur wetland. Submerged macrophytes, on the other hand, provide a more suitable environment for P cycling in the milieu, than the floating macrophytes (Aftabuddin Md., 2013 personal communication). This may explain the significantly higher (p < 0.01) available P in Bhomra. Heavy load of detrital organic carbon in sediments, as recorded by Das (1998) and in the current research can be channelized for fish production through stocking of suitable detritivorous fish species. Alternately, macrophyte coverage in wetlands may be reduced to give space and nutrients for phytoplankton growth. These wetlands are in the process of aging and as a management strategy dredging is occasionally done to remove sediment load and to increase water depth. Re-establishment of river connectivity to these closed wetlands is another proposition to improve wetland health, fish diversity and sustainable production.
This study detected presence of a large organic P pool, indicating its deposition and accumulation in the floodplain wetlands of India. The presence of moderate to high levels of available P in sediment and overlying water emphasizes that sufficient P release is taking place, which can be utilized for phytoplankton growth. Management interventions in terms of partial macrophyte clearance and/or stocking of suitable detritivorous fishes are warranted for effective and economic utilization of the nutrients trapped in these wetlands.
The work was conducted under Indian Council of Agricultural Research funded AMAAS network project on ‘Microbial phosphorus transformations in inland open waters.’
Supplemental data for this article can be accessed on the publisher's website.