James N. Kremer

NITROGEN LOADING FROM COASTAL WATERSHEDS TO RECEIVING ESTUARIES: NEW METHOD AND APPLICATION

In this paper we develop a model to estimate nitrogen loading to watersheds and receiving waters, and then apply the model to gain insight about sources, losses, and transport of nitrogen in groundwater moving through a coastal watershed. The model is developed from data of the Waquoit Bay Land Margin Ecosystems Research project (WBLMER), and from syntheses of published information. The WBLMER nitrogen loading model first estimates inputs by atmospheric deposition, fertilizer use, and wastewater to surfaces of the major types of land use (natural vegetation, turf, agricultural land, residential areas, and impervious surfaces) within the landscape. Then, the model estimates losses of nitrogen in the various compartments of the watershed ecosystem. For atmospheric and fertilizer nitrogen, the model allows losses in vegetation and soils, in the vadose zone, and in the aquifer. For wastewater nitrogen, the model allows losses in septic systems and effluent plumes, and it adds further losses that occur during diffuse transport within aquifers. The calculation of losses is done separately for each major type of land cover, because the processes and loss rates involved differ for different tesserae of the land cover mosaic. If groundwater flows into a freshwater body, the model adds a loss of nitrogen for traversing the freshwater body and then subjects the surviving nitrogen to losses in the aquifer. The WBLMER model is developed for Waquoit Bay, but with inputs for local conditions it is applicable to other rural to suburban watersheds underlain by unconsolidated sandy sediments. Model calculations suggest that the atmosphere contributes 56%, fertilizer 14%, and wastewater 27% of the nitrogen delivered to the surface of the watershed of Waquoit Bay. Losses within the watershed amount to 89% of atmospheric nitrogen, 79% of fertilizer nitrogen, and 65% of wastewater nitrogen. The net result of inputs to the watershed surface and losses within the watershed is that wastewater becomes the largest source (48%) of nitrogen loads to receiving estuaries, followed by atmospheric deposition (30%) and fer- tilizer use (15%). The nitrogen load to estuaries of Waquoit Bay is transported primarily through land parcels covered by residential areas (39%, mainly via wastewater), natural vegetation (21%, by atmospheric deposition), and turf (16%, by atmospheric deposition and fertilizers). Other land covers were involved in lesser throughputs of nitrogen. The model results have implications for management of coastal landscapes and water quality. Most attention should be given to wastewater disposal within the watershed, par- ticularly within 200 m of the shore. Rules regarding setbacks of septic system location relative to shore and nitrogen retention ability of septic systems, will be useful in control of wastewater nitrogen loading. Installation of multiple conventional leaching fields or septic systems in high-flow parcels could be one way to increase nitrogen retention. Control of fertilizer use can help to a modest degree, particularly for optional uses such as lawns situated near shore. Conservation of parcels of accreting natural vegetation should be given high priority, because these environments effectively intercept atmospheric deposition. Areas upgradient from freshwater bodies should be given low priority in plans to control nitrogen loading, because ponds intercept much of the nitrogen transported from upgradient.

A simulated Antarctic ast ice ecosystem

A simple two-dimensional (z,t) model of first year sea ice structure and dynamics is coupled to a high resolution, time-dependent model of microalgal growth in which simulated physiological responses are determined by ambient temperature, spectral irradiance, nutrient concentration, and salinity. The physical component utilizes atmospheric data to simulate congelation ice growth, initial brine entrapment, desalination, and nutrient flux. Temperature gradient, sea ice salinity, brine salinity, and brine volume are also computed. The biological component is based on the concept of a maximum temperature-dependent algal growth rate which is reduced by limitations imposed from insufficient light or nutrients, as well as suboptimal salinity. Estimated gross primary productivity is reduced by respiration and grazing terms. Preliminary simulations indicate that, during a bloom, microalgae are able to maintain their vertical position relative to the lower congelation ice margin and are not incorporated into the crystal matrix as the ice sheet thickens. Model results imply that land fast sea ice contains numerous microhabitats that are functionally distinct based upon the unique suite of processes that control microalgal growth and accumulation within each. In the early stages of the spring bloom, high brine salinity inhibits microalgal growth at all depths within the congelation ice, except near the skeletal layer. Light is predicted to be the limiting resource throughout the congelation ice and platelet ice at this time. Later in the bloom when environmental conditions are more favorable for algal growth, model results suggest that biomass accumulation in the upper congelation ice is controlled by microzooplankton grazing, Microalgae in the skeletal layer and upper platelet ice are susceptible to nutrient limitation at this time due to diminished flux and high nutrient demand. Light limits microalgal growth in the lower platelet ice throughout the bloom. Results indicate that land fast sea ice in McMurdo Sound can support a production rate of approximately 0.5 g C m−2 d−1 under optimal conditions, 76% of which is associated with the platelet layer where rates of nutrient exchange are relatively high. While adjustments in any biological coefficient will alter the magnitude of production in the model, the range of results permitted by uncertainty in their values is well within the bounds likely to result from normal variations in snow cover, or from the uncertainty in the rate of nutrient flux.

A bio-optical model of Antarctic sea ice

Biogenic particulate material in sea ice can substantially influence the spectral irradiance within the ice sheet and underlying seawater. In order to simulate accurately seasonal changes in light conditions in situ, the biomass changes of the sea ice microbial community must be considered. Here we attempt to provide an improved description of the optical regime within sea ice by combining information provided by models of radiative transfer in sea ice and snow and models of solar spectral irradiance with formulations describing the attenuation of spectral irradiance by particulates observed in sea ice in McMurdo Sound, Antarctica. Emphasis has been placed on the role of biogenic particles in visible light attenuation with the intent of developing a bio-optical model that more rigorously describes their influence on radiative transfer processes as they occur in nature. Model results simulating seasonal changes in both photosynthetically active radiation and its spectral distribution agree well with measured under-ice spectral irradiance. Results reveal how changes in microalgal concentrations, as well as their photophysiological characteristics influence both the quantity and quality of downwelled light in sea ice and in the upper layers of the ice-covered oceans.

Estuary-specific variation in the air-water gas exchange coefficient for oxygen

Oxygen air-water gas exchange was measured using floating chambers in two shallow tidal estuaries of differing bathymetry and local terrain, near Waquoit Bay, Massachusetts (United States). The specific chamber design permitted measurements of gas flux in 15 min, allowing analysis of the relationship with wind speed and tidal stage. Exchange coefficients ranged from 0.5 to 2.5 g O2·m−2 h−1 atm−1 (equivalent to piston velocities of 1.5 to 7 cm h−1) for wind speeds of 0.3 to 9 m s−1 at 10 m elevation. While the relationships for each estuary appear linear (significant linear regressions with wind speed were shown for each estuary, and the slopes were different at the 99.5% confidence level), the range of speeds differed at the two sites and an exponential function of wind speed was consistent with the combined data from both estuaries. A power function of wind speed was not an acceptable model. The exchange coefficients for our estuaries are from 57% to as low as 9% of that predicted by previously published generic equations. Because the atmospheric correction can be significant in shallow, metabolically active coastal waters, we suggest that empirically determined relationships for gas exchange versus wind for a specific estuary are preferable to the predictions of the general equations. While the floating chamber method should be used cautiously, at low winds speeds (below 8 m s−1) and in slowly flowing waters, it provides a convenient approach for quantifying these site-specific differences. The differences, especially those between shallow sheltered systems and the open waters best fit by some published relationships, are ecologically important and do not appear yet to be measurable by other methods.

The relative importance of chlorophyll and colored dissolved organic matter (CDOM) to the prediction of the diffuse attenuation coefficient in shallow estuaries

The availability of underwater light is a critical factor in the growth and abundance of primary producers in shallow embayments. The goal of this study was to examine the relative importance of factors influencing light availability in this type of water body. Many simulation models of aquatic ecosystems predict light attenuation from chlorophyll or phytoplankton stock. In the three southern New England sites studied here, no useful relationship was found to exist between chlorophyll and KPAR (the diffuse attenuation coefficient of photosynthetically active radiation; Kirk 1994; Mobley 1994). In 40 of 53 cases, a regression of chlorophyll versus KPAR was not statistically significant. Variation in KPAR did demonstrate a correlation to salinity, implicating a freshwater source of light attenuating material. This was true even in a system with little freshwater inflow. Colored dissolved organic matter (CDOM) is one such terrestrial input that enters estuaries from their watersheds and can strongly influence the availability of light to aquatic primary producers. This study demonstrated that over 70% of the variability in the KPAR coefficient can be attributed to CDOM in the shallow estuaries studied. This illustrates the need for improved model formulations that include CDOM in the prediction of light attenuation in shallow coastal systems. A new equation has been developed to predict KPAR with CDOM.

Volcanic fertilization of Balinese rice paddies

Since the advent of high-yielding ‘‘Green Revolution’’ rice agriculture in the 1970s, Balinese farmers have been advised to supply all the potassium and phosphate needed by rice plants via chemical fertilizers. This policy neglects the contribution of minerals leached from the volcanic soil and transported via irrigation systems. We documented frequent deposition of volcanic ash deposits to rice producing watersheds. Concentrations of phosphorus in rivers were between 1 and 4 mg l − 1 PO4, increasing downstream. We measured extractable potassium and phosphate levels in the soils of unfertilized Balinese rice paddies, and found them to be indistinguishable from those in fertilized paddies, and sufficient for high grain yields. Field experiments varying phosphorus applications to rice fields from 0 to 100 kg superphosphate per hectare (7–26 kg P ha − 1 ) demonstrated small increases in harvest yields only with the smallest additions. Direct measurements of PO4 in irrigation waters indicate that most of the added phosphate flows out of the paddies and into the river systems, accumulating to very high levels before reaching the coast.

Primary production in Long Island sound

Daily and annual integrated rates of primary productivity and community respiration were calculated using physiological parameters measured in oxygen-based photosynthesis-irradiance (P-I) incubations at 8 stations throughout central and western Long Island Sound (cwLIS) during the summer and autumn of 2002 and 2003 and the late spring of 2003. Each calculation takes into account actual variations in incident irradiance over the day and underwater irradiance and standing stock with depth. Annual peak rates, ±95% confidence interval of propagated uncertainty in each measurement, of gross primary production (GPP, 1,730±610 mmol O2 m−2 d−1), community respiration (Rc, 1,660±270 mmol O2 m−2 d−1), and net community production (NCP, 1,160±1,100 mmol O2 m−2 d−1) occurred during summer at the western end of the Sound. Lowest rates of GPP (4±11 mmol O2 m−2 d−1), Rc (−50±300 mmol O2 m−2 d−1), and NCP (−1,250±270 mmol O2 m−2 d−1) occurred during late autumn-early winter at the outer sampled stations. These large ranges in rates of GPP, Rc, and NCP throughout the photic zone of cwLIS are attributed to seasonal and spatial variability. Algal respiration (Ra) was estimated to consume an average of 5% to 52% of GPP, using a literature-based ratio of Ra:Rc. From this range, we established that the estimated Ra accounts for approximately half of GPP, and was used to estimate daily net primary production (NPP), which ranged from 2 to 870 mmol O2 m−2 d−1 throughout cwLIS during the study. Annual NPP averaged 40±8 mol O2 m−2 yr−1 for all sampled stations, which more than doubled along the main axis of the Sound, from 32±14 mol O2 m−2 yr−1 at an eastern station to 82±25 mol O2 m−2 yr−1 at the western-most station. These spatial gradients in productivity parallel nitrogen loads along the main axis of the Sound. Daily integrals of productivity were used to test and formulate a simple, robust biomass-light model for the prediction of phytoplankton production in Long Island Sound, and the slope of the relationship was consistent with reports for other systems.

Potential applications of an empirical phytoplankton production model to shallow water ecosystems

Abstract An existing empirical model for predicting estuarine phytoplankton net production on the basis of chlorophyll a standing crop ( B ), optical depth ( Z p ), and surface irradiance ( I o ) has been adapted for application to shallow water systems (referred to as the BZ p I o model). The original model has proven to be a good predictor of phytoplankton production in estuarine systems where depth exceeds the photic zone. It had yet to be tested in shallow estuarine systems where the photic zone exceeds water depth. The model is relatively simple, with chlorophyll a , light extinction, and daily surface irradiance as its only inputs. Based on existing phytoplankton α and P max data, upon which one of the original BZ p I o models was produced, production across depth was integrated at two irradiance extremes, and related to a third-order polynomial correction formulation ( Z cor ). Due to the tendency for production to be measured via O 2 incubations in shallow systems, this correction term has not yet been tested against 14 C data. O 2 incubations really measure net community production, not net phytoplankton production, and therefore significant variability is introduced. Other factors that may increase variability in this model include measurements and estimates of water column light extinction and depth distribution.

Effects of oyster depuration gear on eelgrass (Zostera marina L.) in a low density aquaculture site in Long Island Sound.

ABSTRACT Oyster (Crassostrea virginica) aquaculture has a long history and tradition in Long Island Sound (Connecticut, USA). Although most of the producers practice traditional on-bottom aquaculture, there are a growing number of individuals utilizing bottom gear for cultivation and depuration. The use of this gear presents a potential conflict in eastern Long Island Sound where the last remaining populations of eelgrass (Zostera marina L.) exist. Shellfish aquaculture activity has been identified as a potential source for negative impacts to eelgrass populations. However, bivalve aquaculture has also been shown to provide an equivalent or greater degree of ecosystem services as submerged aquatic vegetation. The effects of short-term oyster depuration activity were gauged by comparing eelgrass reference sites and experimental plots (eelgrass areas containing oyster depuration cages with and without oysters) in triplicate. Changes in sheath length of the eelgrass 1 m from the cages were used as a proxy for growth rate. The aquaculture gear had no effect on this measure of growth rate of eelgrass in any of the deployments. Sediment characteristics (sediment chlorophyll, sediment % organics) in the cage footprint and 1m from the cages also failed to show an effect of the depuration cages on the local environment. Video monitoring of the footprints and local area indicated little physical damage to the eelgrass beds as a result of the short deployment of the aquaculture gear. The water column at all three sites was vertically well mixed and no effect of the cages on water column light and other characteristics was detectable. The results of this study indicated that at the current level of activity, short-term depuration of oysters has minimal effect on eelgrass growth, water quality and the sediment characteristics measured. However, if depuration activity expands in terms of the amount of gear and/or individual operations, it may result in measurable effects. Understanding the interactions between shellfish aquaculture activity and the marine environment is necessary for sustainable growth of the industry.

Evaluation of ENDECO 1184C dissolved oxygen recorders for use in temperate estuaries

Abstract Pulsed Clarke-type electrodes, a relatively new type of dissolved oxygen (DO) sensor, were intended to minimize problems associated with sensitivity to flow and biofouling-problems which affect most oxygen sensors. The commercially available devices (ENDECO/YSI, Inc., Model 1184C), which include temperature and conductivity sensors and a data logger, are designed for coastal environmental monitoring at depths to 30 m. In order to evaluate their performance under field conditions, a series of field deployments and laboratory calibrations was conducted, and the instrument output was compared with precision Winkler titrations. Accuracy of reported DO concentrations varied both between instruments and with time. Errors typically ranged from 0 to 3 mg/l. Thus frequent calibration seems to be a requirement for the accurate use of these units. The recommended 1-point calibration update did not improve accuracy. In the laboratory, the response to varying DO concentrations at constant temperature was highly linear over a range of approx. 5–14 mg/l; both accuracy and response time decreased at low DO concentrations, somewhere between 0 and 1.5 mg/l. Potential sources of inaccuracy in DO measurements include (1) statistical error associated with the estimation of calibration constants, and (2) operation of the instrument near the extremes of the temperature range used to calibrate it. We found support for the claim that the instruments are relatively insensitive to fouling.