National Estuarine Research Reserve System

System-wide Monitoring Program

Synthesis of the Water Quality Data from 1995 to 2000
Chapter 4: Production, Respiration and Net Ecosystem Metabolism.

Introduction
While no single index has emerged that captures all the complex processes and trophic interactions that occur within estuaries, net ecosystem metabolism is a particularly useful indicator because it integrates the system level processes of primary production and respiration within the estuary. Net ecosystem metabolism may be particularly useful for assessing nutrient enrichment and eutrophication at different locations because this approach provides an index of how well balanced the ecosystem is and appears to reflect the loading of organic matter or dissolved inorganic nutrients to the system (Kemp et al. 1997). Strongly net autotrophic systems like portions of Waquoit Bay (D’Avanzo et al. 1996) and the MERL mesocosms (Oviatt et al. 1986) are dominated by inorganic nitrogen loading, while net heterotrophic systems like Tomales Bay (Smith and Hollibaugh 1993, 1997) are dominated by organic carbon loading. In addition, seasonal changes in metabolic rates, particularly those associated with phytoplankton blooms, can result in changes of net ecosystem metabolism from autotrophy during bloom conditions to heterotrophy during non-bloom conditions (Caffrey et al. 1998).

NERR sites have been selected to be representative of the coastal bioregions of the U.S. and are characterized by a variety of plant communities including phytoplankton, salt marsh, seagrass, mangrove and freshwater macrophyte. Results from the first synthesis (Wenner et al. 2001) demonstrated that metabolic rates were influenced by habitats adjacent to the deployment sites. Sites near mangrove forests and in some salt marsh creeks had exceptionally high respiration rates and were exceedingly heterotrophic. Three sites, located in either eelgrass beds or above macroalgae mats, were autotrophic. Temperature was the single most important factor controlling metabolic rates at individual sites, although salinity was also important at about half the sites. On an annual basis, respiration exceeded gross primary production demonstrating that all but 4 of the 28 NERR sites examined were heterotrophic. Freshwater fill time and nitrogen loading to the different estuaries may explain much of the variance in net ecosystem metabolism.

Data from 42 sites participating in the National Estuarine Research Reserve’s System Wide Monitoring Program (NERR SWMP) between 1995-2000 were analyzed to estimate an integrated index of ecosystem level processes (production, respiration and net ecosystem metabolism). This chapter summarizes how these processes change over a seasonal and annual basis and whether long term trends in the data exist. Relationships among metabolic rates, physical, chemical and biological factors are compared at several Reserves with extensive ancillary nutrient and chlorophyll a data, as well as with literature values from other estuaries.

Methods
Dissolved oxygen (% saturation) data from NERR sites were analyzed following extensive quality control and quality assurance as described in Wenner et al. (2001). For presentation of results, sites were categorized based on the dominant habitat near the deployment site.

Table 14. Land-use, habitat and climate variables used to classify NERR sites.(See PDF for details).

The method used here assumes that water masses are laterally and vertically homogenous (i.e., they have the same metabolic history); thus, in areas where physical processes such as advection and diffusion dominate over biological processes, metabolic rates may be either underestimated or overestimated (Kemp and Boynton 1980). Because previous analyses at two sites (Hudson River, Sawkill site, and Padilla Bay, Joe Leary site) suggested physical forces usually overwhelmed biological activity at these locations (Caffrey 2003), analyses for these two sites were not included in this report. For all other Reserves, data from two sites per Reserve were analyzed.

Oxygen is produced as a by-product of photosynthesis and consumed by respiration. In aquatic environments, oxygen concentrations usually exhibit a characteristic diurnal pattern, with concentrations increasing from morning into mid-afternoon as photosynthesis exceeds respiration. Declining oxygen concentrations occur during the late afternoon or evening in response to decreasing photosynthetic rates and continue to decrease throughout the night when photosynthesis does not occur. In addition to these biological processes, physical processes can also affect oxygen concentrations. Diffusion of oxygen across the air-water interface can increase or decrease water column concentrations, with diffusion from the air into the water occurring when the water is under saturated and vice versa when the water is supersaturated.

The diffusion, or air-sea exchange, was estimated by equation (1) below where (DOsat,t1 DOsat,t2) are the oxygen concentrations (units -%) for t1 and t2 and dt is the time difference, in hours, between t2 and t1. The time interval for all data was 0.5 hours. A coefficient of 0.5g O2 m-2 hr-1 at zero O2 was used to estimate the rate of air-sea exchange (J. Hagy and W.R. Boynton, pers. comm.). The units for air-sea exchange are g O2 m-2. Thus, when the average oxygen concentration for the time interval ((DOsat,t1 +DOsat,t2)/200) is under saturated, air-sea exchange is positive and oxygen diffuses from the air into the water. If oxygen concentrations are supersaturated, air-sea exchange is negative and oxygen diffuses out of the water into the air.

Air-sea exchange =(1-(DOsat,t2DOsat,t1)/200)*.5*dt

This approach may underestimate exchange during periods of high winds and overestimate exchange during calm periods, since previous research has shown that the rate of diffusion is dependent on wind speed (Copeland and Duffer 1964, Hartman and Hammond 1984, Marino and Howarth 1993).

For each time interval, air-sea exchange was subtracted from the change in oxygen concentrations (DO) in g O2 m-2 multiplied by water depth (m) to give oxygen flux (g O2 m-2) as described in equation (2) below.

Oxygen flux = (DOt2 – DOt1)*water depth – air-sea exchange (2)

Oxygen fluxes during daylight hours were combined to give net production. Similarly, oxygen fluxes from night were combined to determine night oxygen flux. Respiration is defined as a positive quantity; thus, night oxygen fluxes were multiplied by “–1” to give a night respiration rate. Gross production and total (day + night) respiration rate were calculated using net production and night respiration values. Assuming a constant respiration during the day and night, night respiration divided by hours of night equals the hourly respiration rate (g O2 m-2 h-1). Total respiration equals the hourly respiration rate multiplied by 24 h (g O2 m-2 d-1). Gross production is the net production plus the respiration occurring during daylight hours and was calculated by adding net production to the hourly respiration multiplied by the daylight hours. Net ecosystem metabolism was calculated by subtracting total respiration from gross production, or more directly by net production minus night respiration. Production rates were converted from oxygen to carbon assuming a photosynthetic quotient of 1.2 (O2:CO2 molar).

Daily metabolic rate data were averaged by month and then by season, defined as winter (December-February), spring (March-May), summer (June-August), and fall (September-November). Means and standard errors of means were calculated for annual data. Seven sites (ACEBB, ELKAP, ELKSM, GRBGB, GRBSQ, NIWOL and NIWTA) had monthly nutrient data for most of the years between 1995 – 2000 and four of these sites (GRBGB, GRBSQ, NIWOL, NIWTA) had monthly chlorophyll a data for the same period. North Inlet sites (NIWTA, NIWOL) also collected dissolved organic carbon (DOC) measurements. Relationships among metabolic rates and temperature, salinity, precipitation, the percent deviation of rainfall from average rainfall, nutrient, chlorophyll a and DOC concentrations were examined using stepwise linear multiple regression analysis. Correlation analysis between annual net ecosystem metabolism and estuarine surface area was performed among habitat groups. All statistical analyses were performed using SYSTAT.

Results
Average annual rates of gross primary production ranged from a low of 2.3 g O2 m-2 d-1 at Old Woman Creek SU to a high of 28.1 g O2 m-2 d-1 at Tijuana River Tidal Linkage (Table 15). The Tidal Linkage site also had the highest total respiration rate as well (32.3 g O2 m-2 d-1) while Elkhorn Slough South Marsh had the lowest respiration rate (4.4 g O2 m-2 d-1, Table 16). Most sites were heterotrophic, with Rookery Bay Blackwater River being the most heterotrophic (7.6 g O2 m-2 d-1). Three sites (Chesapeake Bay VA Goodwin Island, Wells Inlet, and Waquoit Bay Central Basin) were slightly autotrophic.

Spatial and Seasonal Trends by Region
Reserve sites exhibited a strong seasonal pattern of high rates of gross primary production and total respiration in the summer and low rates in the winter, with a few exceptions. Blackbird Landing in Delaware provides a good example of strong seasonal differences (Figure 24). In contrast, seasonal trends at Jobos were very muted with little distinction between winter and summer (Figure 25), perhaps due to the relatively small range in temperature (22-32°C). Net ecosystem metabolism rates also exhibited seasonal patterns, although seasonal patterns for net ecosystem metabolism were weaker than seasonal patterns for production and respiration at most Reserves.(See PDF for details)

Gross primary production “P” and respiration “R” are discussed in the following sections; however, only production plots are shown given consistent P:R trends at all Reserve sites.

West Coast
Reserves in this region span the greatest geographical and climatic gradients in the Reserve system, so it is not surprising that seasonal or inter-annual patterns were not consistent among these Reserves. The South Slough Reserve exhibited the typical summer peak in production and respiration rates, while the Tijuana River Reserve and Bayview Channel (Padilla Bay NERR) had peak rates in the spring (Figure 26a). Elkhorn Slough sites did not show any consistent seasonal trends. Rates were highest at the Tijuana River Tidal Linkage site and least at the Elkhorn Slough South Marsh site (Figure 26a). Elkhorn Slough and Tijuana River did exhibit some inter-annual trends in production and respiration rates, although they were not consistent between the two Reserves. The Reserves in this region were usually most heterotrophic in the summer, except for Padilla Bay (Bayview Channel) and Elkhorn Slough (Azevedo Pond), which were most heterotrophic in the fall (Figure 26b). Padilla Bay (Bayview Channel) was consistently autotrophic in the spring and often autotrophic in the summer.

Northeast
Peak production and respiration rates for all Reserves in this region occurred during summer (Figure 27a). Few Reserves collected data during winter months due to ice cover, but where data were available (e.g. Wells), rates were often near zero. The high winter primary production rate from Great Bay (Squamscott River) should be interpreted with caution because it represents just three days worth of data from a single year. Production and respiration rates were lowest, usually less than 5 gO2 m-2 d-1, at the Old Woman Creek Reserve and the Wells (Head of Tide) site. Three sites in this region (Waquoit Bay, Wells Inlet site and Great Bay GB) were usually balanced or autotrophic with inconsistent seasonal variation. Net ecosystem metabolism at other sites was heterotrophic, often with maximum effects in summer. Inter-annual variation in production and respiration in this region was minimal and inconsistent between sites and among Reserves; however, the Wells Inlet site was somewhat autotrophic in 1996-1997, strongly autotrophic in summer 1998, and then strongly heterotrophic during spring and summer 1999-2000.

Mid-Atlantic
All sites within this region exhibited consistent seasonal trends in production and respiration, except for the Chesapeake Bay Maryland sites where limited sampling makes interpretations difficult. Summer production and respiration rates were often 1.5 to 2 times higher than rates in the other seasons. Production and respiration rates ranged from being relatively low at Mullica River (Lower Bank) to high at both Delaware Bay sites. Interannual variation in production and respiration was minimal in this region. Net ecosystem metabolism showed seasonal and spatial variation among Reserve sites. Maximum heterotrophy during the summer was a consistent pattern across all sites (excluding Chesapeake Bay MD). Chesapeake Bay VA (Goodwin Island) and Mullica River (Buoy 126) sites were usually autotrophic or balanced during the other seasons. Conditions at the Delaware Bay and Mullica River (Lower Bank) sites changed from generally balanced in 1996-1998 to heterotrophic in 1999-2000.(See PDF for details)

Southeast
Production and respiration rates in the Southeast region were greatest at the Sapelo Island sites and decreased moving north into North Carolina (Figure 31a). Most sites had peak rates in summer, except for ACE Basin St Pierre where peak rates occurred in spring. Production and respiration in North Carolina often had a bimodal pattern, with peaks in both March and April and again in June-Aug (Figure 32a). Inter-annual variation in metabolic rates was inconsistent among Reserves in this region. The highest production and respiration rates occurred in 1996 at ACE Basin (Big Bay) and in 1999 at North Inlet (Oyster Landing), but no clear trends were observed for other sites. Similar to the Gulf and Caribbean Reserves, net ecosystem metabolism had a consistent seasonal pattern and was almost always heterotrophic at all sites (Figure 31b). ACE Basin (Big Bay) was the most heterotrophic site in this region, while the North Carolina sites were the least heterotrophic. Net ecosystem metabolism was most heterotrophic during the summer and balanced or slightly autotrophic in the winter, except at ACE Basin (St Pierre) and Sapelo Island (Marsh Landing), which were most heterotophic during spring (Figure 31b). North Carolina sites, particularly Zeke’s Island, were autotrophic or balanced during fall and winter months (Figure 32b). North Inlet (Oyster Landing) also became autotrophic (or balanced) every January or February (data not shown).

Gulf of Mexico and Caribbean
Reserves in the Gulf of Mexico and Caribbean exhibited several different seasonal patterns of production and respiration (Figure 33a). Apalachicola Bay sites and the Weeks Bay-Weeks Bay (WB) site had peak rates in summer, typical of most Reserve sites. Rookery Bay sites and Jobos Bay (site 9) had peak rates in the spring, while rates at both Jobos Bay site 10 and Weeks Bay Fish River showed little seasonal variation (Figure 33a). Metabolic rates were lowest at Apalachicola Bay compared to the other Reserves in this region. Rates within Reserves were generally quite similar, suggesting few differences between control and impact sites. Interannual variation in metabolic rates at all Reserve sites was minimal. In fact, the Blackwater River site in Rookery Bay was autotrophic only 1 day over the entire record. In contrast with the production and respiration rates, net ecosystem metabolism exhibited a consistent seasonal pattern of greater heterotrophy during the summer, except Jobos Bay, which was most heterotrophic during the fall (Figure 33b). All sites from this region were strongly heterotrophic, particularly Rookery Bay sites (Figure 34). (See PDF for details)

Factors controlling metabolic rates
Analysis of the 1996-1998 data indicated that temperature and salinity were important factors controlling metabolic rates at the sites. Nutrient data from some of the Reserves suggested that nutrient concentrations and inputs might also be important in controlling metabolic rates. The seasonally averaged data shown in Figures 26, 27, 29, 31 and 33 confirm the importance of temperature for the entire dataset. Stepwise multiple regression was used to examine how metabolic rates (gross production, respiration, and net ecosystem metabolism) were related to physical (temperature, salinity, rainfall, deviation from normal rainfall), chemical (DIN, DIP, TN, TP, DOC) and biological (chlorophyll a) variables. Regression models for all metabolic rates at the seven Reserves were significant, except for net ecosystem metabolism at the Great Bay GB site (Table 16). The regression models could explain 20-90% of the variation in metabolic rates. R-squared values were generally higher for gross production and respiration than net ecosystem metabolism. Temperature was a significant factor in all the models except for net ecosystem metabolism at both Elkhorn Slough sites. Nutrient concentrations were a significant factor in 5 out of 6 net ecosystem metabolism models, but were significant in only 3 out of 7 gross production or respiration models

While nutrient concentrations are most likely important in controlling metabolic rates, monthly grab samples (or a monthly diurnal set of samples) may not adequately capture water column nutrient concentrations. Even with hourly sampling of nutrients, the relationship between nutrient concentration and metabolic rates are difficult to interpret (Figure 35). Sampling at Elkhorn Slough (Azevedo Pond) provides a good example of this situation. Although higher nutrient concentrations should lead to increased gross production, increased production generally lagged peak nitrate concentrations by two to ten days in January and March (Figure 35). Furthermore, April peaks in gross production appeared to be unrelated to nitrate concentrations (Figure 35).