Variability of Polychaete Secondary Production in Intertidal Creek Networks along a Stream-Order Gradient

Dendritic tidal creek networks are important habitats for sustaining biodiversity and ecosystem functioning in salt marsh wetlands. To evaluate the importance of creek heterogeneity in supporting benthic secondary production, we assess the spatial distribution and secondary production of a representative polychaete species (Dentinephtys glabra) in creek networks along a stream-order gradient in a Yangtze River estuarine marsh. Density, biomass, and secondary production of polychaetes were found to be highest in intermediate order creeks. In high order (3rd and 4th) creeks, the density and biomass of D. glabra were higher in creek edge sites than in creek bottom sites, whereas the reverse was true for low order (1st and 2nd) creeks. Secondary production was highest in 2nd order creeks (559.7 mg AFDM m−2 year−1) and was ca. 2 folds higher than in 1st and 4th order creeks. Top fitting AIC models indicated that the secondary production of D. glabra was mainly associated with geomorphological characters including cross-sectional area and bank slope. This suggests that hydrodynamic forces are essential factors influencing secondary production of macrobenthos in salt marshes. This study emphasizes the importance of microhabitat variability when evaluating secondary production and ecosystem functions.


Introduction
The primary production in salt marshes can exceed 3900 g C m 22 yr 21 , making them among the most productive ecosystems in the world [1]. This high level of primary production forms the energy base for abundant animal consumers, including benthic macrofauna, nekton, and shorebirds, and thereby sustains a high level of secondary production [2,3]. In the salt marshes of the northern Gulf of Mexico, secondary production exceeds other regions of the United States, which is exemplified by large fishery catches of penaeid shrimps (Farfantepenaeus aztecus, F. duorarum and Litopenaeus setiferus of 66% U.S.) and blue crabs (Callinectes sapidus of 25% U.S.) [4]. Similarly, researchers estimated that the secondary production of crabs (Chiromantes dehaani, Helice tientsinensis) in a Yangtze River estuarine salt marsh was 49.4 g AFDM m 22 year 21 , which is higher than that reported from other types of coastal wetlands at a similar latitude [5].
Secondary production is the formation of heterotrophic biomass through time [6]. It represents the cumulative responses of a population to abiotic and biotic stresses, and is an essential variable when quantifying food webs and ecosystem functioning such as material cycling and energy flow [7]. The habitat heterogeneity has been noted when estimating secondary productions. The secondary production of macrobenthos was highest in sandy organic sediment and lowest in muddy sediment in a southern New England salt marsh [8]. Secondary production of macrobenthos differed with vegetation types in the Mobile-Tensaw Delta [9]. A better understanding of the spatial differences in production, especially at local scales, would be helpful to assess the spatial variability of ecosystem functioning.
Intertidal creeks are an important sub-habitat in salt marsh ecosystems and are used by shorebirds and benthic, nektonic, and planktonic animals [10]. These habitats play a key role in promoting material cycle and energy flow between the salt marsh and offshore ecosystem [11]. According to the stream order classification, the smallest headwater stream is defined as a 1st order creek, and creek order increases thereafter when two creeks of the same order connect [12]. Geomorphological characters (e.g. length, width, and depth) and physicochemical properties (e.g. velocity, water temperature and substrate) are usually different between creeks of different orders [13,14,15]. Thus, creek order could be an important factor for evaluating the effects of local habitat heterogeneity on secondary production.
Polychaetes are the dominant macrobenthic component in many tidal creek ecosystems [16]. Their density can exceed 14,500 ind m 22 and serve as an important food source for higher trophic level consumers [17,18]. They are also the prominent ecosystem engineers in salt marsh ecosystems. They can affect the organic matter content in sediment via bioturbation [19], accelerate the erosion of intertidal creeks [20], and stimulate the expansion of some plant species [21]. An understanding of the spatial variability in secondary production of this functionally important organism is essential for determining the processes that affect benthic ecosystems across different habitats. As to polychaete, the highest secondary production of Caulleriella caputesocis was found in the sediment containing 60 to 100% silt [22]. In the Río de la Plata estuary, secondary production of Laeonereis culveri was higher in creeks than in intertidal flat [17]. There are various microhabitats within tidal creek networks, but little is known about the spatial variability in polychaete production.
Dentinephtys glabra (Hartman), which belongs to the family Nephtyidae, is the most numerically dominant macrofaunal species in the creek systems of the Yangtze River estuarine marshes [23,24]. D. glabra is a typical infaunal species which has low mobility and is considered an ideal and representative species to study the effects of sediment environmental variability [25]. This species is also a key link between primary producers and higher trophic levels such as nektonic fish and shorebirds, and thus plays an important role in the food web of the marsh ecosystem. To unravel the importance of creek heterogeneity on macrobenthos, we selected this representative polychaete species to evaluate the effects of creek order on benthic population and secondary production. We aim to answer the question how the distribution and secondary production of D. glabra differs among creeks of different orders.

Ethics statement
This study was conducted according to the regulations of local and central government. Sample collections necessary to scientific research were authorized by the Specifications for oceanographic survey -Part 6: Marine biological survey (GB/T 1263. 6 -2007).

Study site
The study site was located in Shanghai Jiuduansha Wetland National Nature Reserve (31u039-31u179N, 121u469-122u159E) and covered a total area of 420.2 km 2 ( Figure 1). Jiuduansha, with a history of about only 60 years, is a young intertidal island in the Yangtze River estuary. It was formed mainly due to the connection of flood tidal currents and ebb tidal currents. Since its formation, it has experienced considerable enlargement, heightening and shape modification because of the dynamic tidal fluxes in the estuary. The salt marsh is dominated by the native plants (Phragmites australis and Scirpus mariqueter) and an exotic plant (Spartina alterniflora). The mean annual temperature, precipitation, and salinity is 15.7uC, 1145 mm, and 11.7, respectively. The intertidal zones are flooded by irregular semi-diurnal tide with the tidal range from 2.45 m to 4.96 m (mean tidal range 2.75) [26].

Sampling design
Monthly samples were collected from a four order creek system during the neap tides between May 2008 to April 2009. We sampled three replicate creeks per stream order ( Figure 1). In each creek, we sampled at two locations, the deepest point of the creek (bottom), and an edge position approximately 1 m below the high tide level (edge). All the samples were taken in the intertidal zone and not within the vegetation. At each location, we collected five cores (176.6 cm 2 in area and 20 cm depth for each core) at intervals of 2-3 m. Sediments from these five cores were mixed to form a composite sample for polychaete collection. Thus, we collected a total of 288 composite samples during the study. The sediment samples were washed through a 0.5 mm mesh sieve. D. glabra were sorted and fixed in 10% formalin. The top 5 mm of sediment within a 468 cm area was collected at each location and preserved at 220uC for chlorophyll a (chl a) concentration analysis [27]. Additional sediment samples down to a depth of 10 cm (19.6 cm 2 ) were taken for analysis of water content and grain size. Sediment physical parameters (temperature, pH, and conductivity) were directly measured during each month, except May, June, and July, 2008. A global positioning system (GPS) was used to ensure we sampled the same location each month.

Laboratory procedures
Sediment was dried to constant weight at 80uC to measure water content. Chl a content in the sediment was determined using spectrophotometry to represent benthic algae biomass [28]. Algae are an important food source for many primary benthic consumers. Although D. glabra do not ingest algae directly, there are indirect relationships between algal biomass and the carnivorous consumer species. Sediment samples for grain size analysis collected from the same location in June, September, and December 2008 and March 2009 were mixed and measured using a LS-POP (VI) laser grain sizer. Geomorphological characters included cross-sectional area, bank slope, bank full elevation, width and depth with the detail measuring methods followed our previous study in [29].
We used this width at setiger 8 (Wd8S) as an indicator of the total length of the polychaetes [30]. The Wd8S of each D. glabra individual was measured using a photo imaging system (NIS-Elements D 2.20) to the nearest 0.01 mm. Complete specimens of D. glabra were dried at 60uC until constant weight to estimate dry mass (DM). The ash free dry weight (AFDM) was obtained by incinerating at 450uC in a muffle furnace for 8 h.

Data analysis
We used the relationship between AFDM and Wd8S (AFDM = 0.8082 Wd8S 1.1226 , N = 32, R 2 = 0.9792) to estimate biomass and secondary production. To estimate secondary production, species were grouped into several size classes at 0.4 mm intervals based on Wd8S.
Secondary production was calculated using the size frequency method [31,32], expressed as: where P is the secondary production (g AFDM m 22 year 21 ), i is the number of size classes, n j is the mean density in size class j (ind m 22 ), W j is the mean individual weight in size class j (g AFDM), and CPI is the longevity time from hatching to death of the largest size class. The seasonal VBGF (von Bertalanffy growth function) was used to estimate the longevity of D. glabra [33]: where L t is the length (Wd8S) at time t (mm), L ' is the theoretical maximum length (Wd8S), K is the curvature parameter, C is a constant representing the amplitude of seasonal growth oscillation, t 0 is the age at zero length (Wd8S), wp is the period when growth is slowest. The value for t 0 was estimated on the basis of the smallest width found [30]. Life span was calculated based on the age at which 90% of L ' was achieved [17]. Because creeks of the same order have similar width [29], the creek area of each order was calculated based on creek width and its total length. We used ArcGIS 9.3 to measure the length of each creek on Jiuduansha Island based on satellite image data of 2 m resolution.
One-way ANOVA was used to evaluate the effect of creek order on environmental parameters. Two-way or one-way ANOVA was used to evaluate the effects of creek order and sampling location on the density and biomass of D. glabra. The effect of creek order on secondary production of D. glabra was also evaluated by oneway ANOVA. Post hoc multiple comparisons were performed using the least square differences (LSD) test. Differences were regarded as significant at P,0.05. To meet the assumptions of ANOVA, numeric data were log (x+1) transformed before statistical analyses when necessary.
To examine the relationship between secondary production of D. glabra and environmental factors, we performed model selection in regression analyses using Akaike's information criterion (AIC) corrected for small sample size (AIC C ) [34]. Spearman correlation were used to exclude the auto-correlated factors with the absolute coefficients larger than 0.7 [35]. Water content, Chl a concentration, conductivity, temperature, pH, clay-silt content, bank full elevation, cross-sectional area and bank slope were considered in models. Creek order was a categorical factor. Numeric data were log transformed before analyses. We calculated the Akaike weight (w i ) and R 2 for each model. Akaike weight (w i ) represents the probability that the model i is the best model among the given set of alternatives [34]. The proportion of variance explained by the model (R 2 ) was calculated for each model [36].

Results
Environmental factors by creeks and orders are showed in Table 1. In creeks across 1st to 4th orders, the ranges of sediment water content, Chl a concentration, pH, conductivity, temperature and clay-silt content were 52.60-78.91 (%), 0.02-30.60 (mg/g), 6.39-8.55, 1.52-7.97 (ms/s), 0.55-30.95 (uC) and 65.79-87.31 (%), respectively ( Table 1). The highest water content, temperature, pH and sand content were in the 4th order creek. The highest Chl a concentration and clay-silt content were in 2nd and 1st order creek, respectively. The values of bank full depth, width, and cross-sectional area increased with increasing creek order.
The density and biomass of D. glabra are shown in Figure 2 (Figure 2). The 3rd order creek had significantly higher densities (P = 0.0002) and biomass (P = 0.0013) of D. glabra than other order creeks. This spatial pattern varied with different sampling months. From May to August, the highest density and biomass of D. glabra occurred in the 2nd order creeks, whereas between September to April levels were highest in the 3rd order creeks (Figure 3).
For the 1st and 2nd order creeks, the density (1st: P = 0.0222; 2nd: P = 0.0160) and biomass (1st: P = 0.0331; 2nd: P = 0.0118) of D. glabra were significantly higher at the bottom locations than at the edge locations ( Figure 4). However, the trend was reversed in     The secondary production of D. glabra is shown in Table 2. A total of 6222 individuals were collected during the year-round sampling with the mean Wd8S ranging from 0.84 to 1.59 mm. Using the size frequency method (Table S1- (Table 2). Secondary production differed among creeks of different orders. The secondary production in the 2nd order creeks was 2.26 times greater than in the 4th order creeks.
The secondary production of creek system in Jiuduansha marsh is shown in table 2. The creek system in Jiuduansha marsh was 2952.0610 3 m 2 in area, with the 1st order creeks accounting for 51.1% of the total area (1507.9610 3 m 2 ). The average and total production of D. glabra in creeks of the Jiuduansha marsh was estimated to be 387.2 mg AFDM m 22 year 21 and 1143.8610 6 mg AFDM year 21 .
The top five selected AIC models for the relationship between secondary production of D. glabra and environmental factors are shown in Table 3. During the analyses, the cross-sectional area was selected as an integrated performance for bank full depth and width. Clay-silt content was selected to represent grain size of the sediment. All five models had the similar w i value, and included cross-sectional area, bank slope and Chl a concentration, suggesting the importance of these variables.

Spatial distribution
Our results showed that the highest density and biomass of D. glabra occurred in intermediate order creeks. The density of the polychaete Streblospio benedicti has been reported to increase with increasing creek order because the creeks of higher order had more constant physical characteristics [37]. Interestingly, the inverse trend of higher polychaete densities in lower order creeks than in higher order creeks has also been documented [11,24,38]. Taken together, these observations suggest that the macrofaunal utilization pattern of creeks along an order gradient differs with species or regions.
Researchers have found that the density of polychaetes was higher in the edge sites than in the bottom sites [23]. Interestingly, the densities and biomass of D. glabra were higher in edge locations than in bottom locations in the high order (3rd and 4th order) creeks, and lowest in the edge locations of low order (1st and 2nd order) creeks. Essential abiotic properties including sediment grain size and tidal forces may influence the macrobenthos distributions [39]. For example, the density of polychaete Capitella spp. was positively correlated with percent silt [40]. In high order creeks, macrobenthos may be exposed to more intense hydrodynamic  Table 2. Estimation of the total Dentinephtys glabra production in Jiuduansha creeks on an order by order basis with one-way ANOVA results showing the effect of creek order.
Creek order Dentinephtys glabra  Table 3. Results of model selection using Akaike's information criterion (AIC) for the relationship between secondary production of Dentinephtys glabra and environmental factors. forces in the bottom sites than in edge sites, resulting in lower abundance at bottom sites [23].
The pattern of polychaete distribution may also be influenced by other organisms. For example, the presence of burrowing crabs (Chasmagnathus granulatus) reduces the abundance of polychaetes [41]. This is because the crabs reduce the burrowing depth of polychaetes, thereby increasing the availability of the polychaetes to shorebirds [41]. In salt marsh creeks of the Yangtze River estuary, the abundance of the crab Ilyoplax deschampsi was higher in edge sites than in bottom sites [42]. The effects of I. deschampsi on D. glabra may be more evident in lower order creeks where I. deschampsi is more abundant than in higher order creeks [24]. This may also explain the difference in D. glabra abundance between edge and bottom sites between the higher and lower order creeks. Predator is also an important factor to influence the secondary production of polychaete [43]. The secondary production of D. glabra was at least two fold higher in intermediate order creeks than in both low and high order creeks. In high order creeks, longer tidal inundation exposed D. glabra to carnivorous fishes for longer time than in low order creeks. However, in the low order creeks, benthic polychaete D. glabra would be more exposed to shorebirds because of limited tidal inundation [44]. The pattern that higher secondary production of D. glabra occurred in intermediate order creeks may reflect the predatory stress to some extent.

Secondary production
Studies evaluating the secondary production of Nephtyidae are very rare, in fact only three species (Nephtys hombergii, N. australiensis and N. picta) have been studied to date. The secondary production of N. australiensis was 48.0 mg AFDW m 22 year 21 [45]. The secondary production of N. hombergii ranged from 92.0 [46] to 7335.0 mg AFDW m 22 year 21 [47], and that of N. picta from 95.0 to 628.0 mg AFDW m 22 year 21 [48]. Compared with these reports, secondary production obtained for D. glabra in this study was within the range of 48.0 to 7335.0 mg AFDM m 22 year 21 .
Tidal influences result in wide fluctuations of environmental variables within intertidal creek networks [37]. Turbidity and water temperature decreased, while pH, conductivity, flow, and drainage basin area increased as creek order increased [14]. In our study, Chl a concentration and sediment grain size were found to be different among different creek orders. Secondary production is influenced by many physicochemical factors, e.g. temperature, dissolved oxygen, substrate [49]. Secondary production of the polychaete N. hombergii differed among areas of different organic matter content [44]. Research found the decrease of secondary production of polychaete Nereis diversicolor may be attributed to the increase of salinity and the decrease of organic matter content in the estuary of Oued Souss [50]. In the top fitting AIC models, secondary production of D. glabra was mainly associated with geomorphological characters including cross-sectional area and bank slope. This may reflect that the hydrodynamic forces are essential factors for secondary production of macrobenthos in salt marshes. Animals are not evenly distributed across wetland landscapes and the distributions of secondary production change with the high complex landscape topography [51]. Recently, researchers have incorporated the use of GIS analysis to estimate the secondary production for a region. For example, they calculated the area of each sub-habitat (different distances from the marsh edge) using GIS and then calculated the total secondary production of crustaceans for the Galveston Bay salt marsh [52]. In the present study, because of the difference in D. glabra production among different order creeks, we think it is necessary to combine area estimation using GIS analysis to estimate total secondary production in the Jiuduansha marsh. Based on this approach, the total production of D. glabra in Jiuduansha was 1143.8610 6 mg AFDM year 21 . We believe that improved knowledge of the spatial heterogeneity in secondary production combined with landscape ecology is critical to quantify the material cycling and energy flow for the creek networks.

Conclusions
The major findings of this study were: (1) the density and biomass of the polychaete D. glabra differed significantly among creeks of different orders, but was highest in intermediate order creeks. The highest density and biomass were found in the 3rd order creeks and the lowest were in the 4th order creeks. (2) In lower order creeks, its density and biomass were higher in the creek bottom than in creek edge sites. However, the reverse was true in higher order creeks. (3) The secondary production of D. glabra was much higher in intermediate order creeks than in low and high order creeks. Its secondary production in the 2nd order creeks (559.7 mg AFDM m 22 year 21 ) was 2.26 times greater than in 4th order creeks.
Polychaetes are a dominant group and key linkage in the food web of salt marsh ecosystems. The research about polychaete secondary production can be an indicator to evaluate the health of the ecosystems. Our findings indicated the importance of creek order in influencing the abundance and secondary production of macrobenthic animals such as polychaetes. This suggests that future studies of infauna on tidal creek systems should carefully consider the within-system differences. The spatial variability in prey production can, in turn, influence the habitat utilization by predators including nektonic fish and crustaceans, and birds. Thus, the value of different microhabitats should be considered in assessing marsh functions and implementing management and restoration strategies for the purposes of biodiversity conservation.

Supporting Information
Table S1 Annual production of Dentinephtys glabra at 1st order creeks estimated by the size-frequency method. (DOC)