Time after flood (d)

FIGURE 11.9 Decline of nitrate uptake length as the periphyton mat and associated organisms recovered following a large flood in Sycamore Creek, Arizona, estimated from natural declines in nitrate concentrations in streamwater. (Reproduced from Marti et al. 1997.)

all combined ammonium uptake on the stream bottom), and secondarily to nitrification (2030%) (Peterson et al. 2001, Webster et al. 2003). Rates of uptake and nitrification varied as widely among forested New Hampshire streams as was observed in the cross-biome study of Peterson et al. (Bernhardt et al. 2002). Variation in stream hydraulic features and availability of organic C likely are responsible for much of this variation, and there is some evidence that rates of nitrification may be reduced when inputs of nitrate are high.

Biotic demand has been shown to influence removal of phosphate, particularly during periods when streams receive fresh inputs of leaf litter. Phosphorus uptake lengths were negatively correlated with amount of leaf litter in a woodland stream, reflecting microbial demand associated with decaying leaves (Mulholland et al. 1985a). When sediment microbes were killed using chlorine in streamside artificial channels, uptake lengths more than doubled, confirming that biotic uptake influenced spiraling distance (D'Angelo et al. 1991). When leaf litter and then wood was removed from an Appalachian headwater stream, nutrient uptake lengths increased substantially in comparison to a reference stream (Webster et al. 2000). A comparison of two streams that differed in community respiration (CR) found a higher demand for P in the stream with the higher CR (Mulholland et al. 1997).

Hydrology is a key environmental variable affecting nutrient dynamics in streams, influencing inputs, the magnitude of transient storage, and, acting through velocity and discharge, the abundance of primary producers and the retention of particulate organic matter. Collectively these effects of streamflow determine whether the stream is shifted toward throughput or processing mode as the Sycamore Creek example illustrates. In Antarctic streams, where algal mats dominated by filamentous cyanobacteria decrease during periods of extreme low flows as well as in response to scouring high flows, nutri ent retention varies accordingly (McKnight et al. 2004). Other environmental variables can limit primary producers (Section 6.1.1), of course, and thus also reduce nutrient uptake. In a comparison of forested versus logged reaches of a Mediterranean stream, Sabater et al. (2000) found greater algal biomass in open relative to shaded reaches, and uptake lengths for ammonium and P also were shorter in the open reaches. Retention efficiency was greater overall for ammonium than for phosphate, but the latter was more affected by shading (Figure 11.10). The uptake velocity for P was correlated to primary production in both reaches, suggesting that its retention was due to algae. However, vf for ammonium was poorly correlated with algal measures, indicating that microbial heterotrophic

2 Jun 23 Jun 11 Jul 26 Jul 8 Sep 29 Sep

FIGURE 11.10 Temporal variation in ammonium and phosphate mass transfer coefficients (vf) in logged and shaded stream reaches. (Reproduced from Sabater et al. 2000.)

2 Jun 23 Jun 11 Jul 26 Jul 8 Sep 29 Sep


FIGURE 11.10 Temporal variation in ammonium and phosphate mass transfer coefficients (vf) in logged and shaded stream reaches. (Reproduced from Sabater et al. 2000.)

processes or abiotic mechanisms were responsible. Environmental factors that influence algal and microbial production often vary seasonally, leading to changes in nutrient uptake and retention. In two tussock grassland streams in New Zealand, uptake velocity was highest during the spring and lowest during fall and early winter, which corresponded to changes in algal abundance (Simon et al. 2005).

Aquatic macrophytes and bryophytes are capable of removing substantial amounts of nutrients from flowing water. Meyer (1979) recorded significant removal of P as a pulse passed over a bryophyte bed in a forested stream in New Hampshire. In Walker Branch, Tennessee, ammonium uptake by the bryophyte Porella represented 41% of total N retention at the end of a 6-week 15N addition experiment (Mulholland et al. 2000). Although laboratory studies show that rooted aquatic plants reduce pore-water nutrient pools by their metabolic uptake, field studies in the middle Hudson River showed either no such reduction due to submerged aquatic vegetation, or enrichment of nutrient porewater pools (Wigand et al. 2001). The explanation for this apparent contradiction is thought to be the accumulation of allochthonous particulate organic matter in vegetation beds and its mineralization by microbes, which replenishes pore water concentrations. In effect, the uptake and sequestration of nutrients by macrophyte beds is masked by processes that promote replenishment of porewater nutrients, thereby enhancing nutrient retention in the river system (Wigand et al. 2001). A similar effect has been observed in the Spree River, Germany, where macrophyte beds slowed currents and increased water residence time, enhancing organic matter deposition and contributing significantly to monthly P retention (Schulz et al. 2003).

Seasonality in biological activity has been shown to impose seasonal trends on nutrients. In a forested headwater stream in Tennessee, streamwater nitrate and phosphate concentra tions were lowest in autumn and early spring, corresponding to periods of highest heterotro-phic and autotrophic activity, respectively (Mulholland 2004). High uptake rates are seasonally driven by microbes colonizing fresh leaf litter and the photosynthetic demands of algae and bryophytes, and are reduced by winter discharges that flush organic matter and by shading due to summer leaf-out, respectively. In this system, instream processes can remove up to 20% of the nitrate and 30% of the phosphate that enters the stream annually (Mulholland 2004).

Nutrient uptake rates generally increase with increases in nutrient concentrations until supply exceeds biotic demand, at which point saturation of nutrient uptake occurs (Dodds et al. 2002, Simon et al. 2005). In agricultural streams of the Midwestern US where nutrient concentrations were 10-100 times greater than in pristine systems, nitrate uptake appeared to be saturated because uptake was not higher at the more enriched sites and vf declined at sites with greater concentrations. Uptake of P and ammonium were not saturated because both increased with higher water column concentrations, but declining vf at high concentrations suggests that both may have been approaching saturation (Bernot et al. 2006). At the highly enriched site below a sewage treatment plant in Arkansas, described previously, biological uptake had no discernible effect on P concentrations; instead, buffering by abiotic uptake and release mechanisms stabilized streamwater concentrations (Haggard et al. 2005). Dissimilatory transformations

By transforming ammonium into nitrate, nitrifying bacteria can influence the concentrations of each form of N in streamwater as well as the possibility of its eventual permanent removal through denitrification to nonreactive N2. Nitrification accounted for over half of total ammonium removal in Eagle Creek, Michigan (Hamilton et al. 2001), and in Quebrada Bisley, Puerto Rico

(Merriam et al. 2002), but for <20% in Walker Branch, Tennessee (Mulholland et al. 2000). The resultant nitrate can be immobilized by the biota, exported downstream, or transformed into N2 gas by denitrifying bacteria. Nitrification was implicated as a source of nitrate in a Mojave desert stream, where streamwater nitrate concentrations were two times greater than expected from groundwater inputs (Jones 2002). Mineralization of organic N to ammonium followed by nitrification to produce nitrate was thought to be responsible for the additional nitrate in streamwater. Because nitrifying bacteria require aerobic conditions, nitrification rates should be greater where subsurface flows provide oxygenated water, such as in shallow, disturbed sediments. Nitrification rates within the hyporheos of Sycamore Creek were highest in regions of downwelling, which presumably supplied organic C and oxygenated water (Jones et al. 1995). In this system, most of the N demand of surface algae was met by hyporheic mineralization of ammonium and subsequent nitrification, demonstrating the importance of the coupling of surface and subsurface systems.

The reduction of nitrate to nonreactive N2 gas by denitrifying bacteria is an important pathway in the N cycle because it represents the only process that permanently removes reactive N from the stream network. Denitrification rates are enhanced under conditions of high nitrate and organic matter availability and low oxygen concentrations. Denitrifying bacteria can be found in association with FBOM and senescing mats of Cladophora and cyanobacteria, where anoxic zones develop that bacteria can utilize (Triska and Oremland 1981, Kemp and Dodds 2002a, b). Sediment-rich streams in agricultural watersheds exhibit high denitrification rates (David and Gentry 2000, Royer et al. 2004), but nitrate levels nonetheless remain high owing to very high inputs of nitrate from fertilizer.

Denitrification rates tend to be higher under conditions of low flows and shallow depths, owing to greater interaction of water with sediments, and so the smaller and shallower lower-order tributaries likely play a disproportionate role in nitrate removal (Alexander et al. 2000). In a headwater stream in Tennessee, de-nitrification accounted for 16% of total N uptake (Mulholland et al. 2004), and estimates from individual stream reaches generally find N removal attributed to denitrification to be <20% (Seitzinger et al. 2002). Based on a review of published studies of N loss determined by direct measurements and mass balance estimates, Seit-zinger et al. conclude that the amount of N lost from all streams of a catchment is much greater than the fraction removed from a single reach, because of the cumulative effect of N removal along the entire flow path. For their model constructed for rivers of the Northeastern US, from 37% to 76% of N inputs were lost by denitrifica-tion along the river network, about half in first through fourth-order streams, and the remainder in higher-order rivers. This differs from the findings of Royer et al. (2004) reported above, who suggest that flood plains and wetlands may be more important than headwater streams as locations of denitrification in the agricultural Midwest.

The measurement of denitrification rates is an area of continuing methodological improvement, but still faces measurement challenges. Direct measurement requires field chambers to enclose sediment cores or transport of sediment slurries to the laboratory, both of which may introduce artifacts (Martin et al. 2001). Recently developed 15N tracer techniques (Rysgaard et al. 1993) applied to whole stream reaches (Mulholland et al. 2004) allow direct measurement of nitrate uptake and denitrification rates by quantifying both the removal of 15NO3-N from the water column and the direct labeling of 15N2 and 15N2O gases. Although the loss of nitrate from streamwater commonly is attributed to denitrification, dissimilatory reduction of nitrate to ammonium (DNRA) has recently been proposed as an alternative pathway. Because the end product is ammonium rather than N2

gas, this process does not result in permanent removal of N from the stream system (Whitmire and Hamilton 2005). Role of consumers

The animal community directly influences nutrient cycling by consumption and assimilation of algal and bacterial production, and by the subsequent release of nutrients by egestion of feces and excretion of urine. By consuming lower trophic levels, especially primary producers and heterotrophic microorganisms, animal consumers enhance the rate of regeneration of nutrients by excretion and egestion, thus replenishing the pool of available inorganic nutrients. Most excreted nutrients are inorganic forms such as ammonium and phosphate, and thus are available to primary producers, but some nutrients such as urea are excreted as organic forms. Egested materials also are in organic forms and so must undergo decomposition and mineralization by microbes to be available for uptake and assimilation. Because nutrient availability can strongly limit overall productivity of a stream ecosystem, high rates of consumption and secondary production can stimulate nutrient cycling and thus help to maintain a system in a highly productive state.

Sycamore Creek, Arizona, is an example of a stream that sustains very high rates of secondary production despite being strongly N limited (Grimm and Fisher 1986), evidence of a high rate of nutrient regeneration and probably of reingestion of feces (Fisher and Gray 1983). Based on laboratory estimates of mass-specific excretion and egestion rates, Grimm (1988) estimated that up to one third of ingested N was converted to ammonium and so was readily available to autotrophs. About half of ingested N was egested as fecal material and presumably was recycled via reingestion by consumers, or became available for uptake following leaching or micro-bial breakdown. Although Grimm was unable to determine the exact amount of recycled N, which she estimated to fall between 15% and 70% of whole stream N retention, even the lower value implies a significant role for animals in nutrient regeneration in this highly productive, N-limited system. Consumer excretion was also shown to make a major contribution to the nutrient demand in a highly productive geothermal spring stream in Yellowstone National Park, with warm temperatures throughout the year and high primary productivity due to abundant filamentous algae and vascular plants (Hall et al. 2003). The exotic snail Potamopyrgus antipodarum consumed 75% of algal production and its excretion supplied two thirds of the algal mat's ammonium demand. High snail biomass rather than high per-biomass rates of consumption and excretion explained the snail's dominant role in nutrient flux in this atypical system.

Retention versus release of nutrients by consumers is influenced by body size, temperature, and the consumer's metabolic demand. The amount of nutrients excreted per unit body mass and time usually decreases with increasing body mass as a consequence of the scaling of metabolism with body size (Vanni 2002). Excretion rates increase with temperature due to the influence of temperature on metabolic rate. For animals to maintain relatively constant nutrient content of their body mass, they will incorporate nutrients at a rate necessary to meet their metabolic needs and excrete nutrients not needed for growth (Sterner and Elser 2002, Cross et al. 2005). Ecological stoichiometry theory further asserts that animals feeding on a nutrient-poor diet will retain more nutrients relative to those feeding on a nutrient-rich diet, and the ratio of nutrients retained versus excreted should reflect any imbalance between the nutrient requirements of the consumer and the nutrient ratio in the food supply. In general, the N/P ratio released by an animal should be negatively correlated with that of its body mass and positively correlated with that of its food (Vanni 2002). A comparison of excretion rates and ratios for N and P for 28 species of fishes and amphibians in Río las Marias, an Andean piedmont stream in Venezuela, supports expectations from stoichi-ometry theory (Vanni et al. 2002). Phosphorus excretion rates and N/P excretion ratios were negatively correlated with body P content and body N/P ratios, respectively (Figure 11.11). Total excretion by the assemblage of consumers in this stream was estimated to meet 49% of algal demand for N and 126% of algal demand for P. Interestingly, armored catfish have a very high body content of P, consume a stoichiometrically imbalanced diet that is low in P, and produce excretions of low P content and high N/P ratios (Hood et al. 2005). As a result, these herbivorous fishes appear to act as P sinks, decreasing the availability of P for algae, which in turn may result in lowering the quality of food available to consumers.

Animals also transport nutrients among habitats and across ecosystems by their movements and migrations. Emergence of the adult stages of aquatic insects is one such process but various authors agree that it is a small fraction of overall nutrient transport (<1%, Meyer et al. 1981,

FIGURE 11.11 Relationship between phosphorus body content and excretion rate in fishes and amphibians in Río las Marias, Venezuela. (Reproduced from Vanni et al. 2002.)

Grimm 1987, Naiman and Melillo 1984, Triska et al. 1984), although potentially important to animal consumers of the riparian zone as a supply of organic C (Section 10.4.1). In contrast, spawning runs of anadromous fish may import substantial amounts of marine-derived nutrients to streams and lakes by their excretion, release of gametes, and their own mortality, especially if many or all die after reproducing. Spawning salmon provide an important nutrient subsidy to freshwater ecosystems of the Pacific Northwest that are generally nutrient-poor (Naiman et al. 2002). Where salmon are abundant, a large proportion of the N in the stream biota likely is derived from spawning fish (Bilby et al. 2001), and significant quantities appear in riparian vegetation and a host of animal consumers. In Sashin Creek, Alaska, isotope analysis showed that N and C derived from a spawning run of Pacific salmon were incorporated into periphy-ton, macroinvertebrates, and fish (Kline et al. 1990). In a wine-growing region of California, cultivated grapes adjacent to Chinook salmon spawning sites obtained as much as 25% of their N from marine sources (Merz and Moyle 2006). A comparison of an Alaskan stream in which downstream reaches contained salmon but upstream reaches did not, found that salmon-supporting reaches had higher SRP concentrations, epilithon abundance, and chironomid biomass but mayfly biomass was lower (Chal-oner et al. 2004). Because of declines in salmon populations, it is estimated that only 6-7% of the marine-derived N and P that historically was transported into the rivers of the Pacific Northwest by salmon is currently reaching those rivers (Gresh et al. 2000).

Animal consumers also affect nutrient cycling indirectly, through their influence on benthic algal biomass, organic matter dynamics, and prey assemblages (Vanni 2002). Whenever grazing sharply reduces algal and biofilm biomass, uptake rates are expected to decrease and spiraling distance to lengthen, whereas moderate grazing that stimulates primary and microbial production should have the opposite effect. Consumption of algae and leaf litter by the snail Elimia clavaeformes in artificial channels stimulated mass-specific metabolic rates of microbial populations, but periphyton and microbial biomass were so reduced that overall biotic uptake of P declined (Mulholland et al. 1983, 1985a). As a result, spiraling distance was shortest in the absence of snails. Low grazing pressure by snails allowed the accumulation of an algal mat that created transient storage effects, which in turn enhanced nutrient recycling within the mat (Mulholland et al. 1994). The conversion of CPOM into FPOM by detriti-vores is likely to enhance transport of particulate N and P because fine particulates are more easily exported during storms, but may also favor retention whenever fine particles are consumed (Vanni 2002). Predators may affect nutrient regeneration through their influence on the prey size spectrum because mass-specific nutrient excretion declines with body size. If larger prey are preferentially ingested, the excretion rate of the remaining, smaller prey assemblage will increase (Vanni 2002).

0 0

Post a comment