Minutes of Coastal Bays Nutrient Budget Workshop

Date:    Nov 22, 2004

Place:   Tawes State Office Building


In Attendance:

Walter Boynton



Tom Jones


Tom Jordan



Liliana Gonzalla


Bruce Michael



Lora Harris


Michael Luisi



Luke Cole


Jim Casey



Jon Dillow


Jim George



Bob Shedlock


John Sherwell



Steve Doctor


Cathy Wazniak



Jane Thomas


Tom Parham



Matt Hall


Margaret McGinty



Eva Bailey


Brian Sturgis



Bill Dennison


Mike Owens



Darlene Wells


Ken Shanks



Jason Dubow


Carol Cain



Roman Jesien


Jeff Cornwell






Perspectives of N and P loading:  Walter Boynton

Nutrient budgets are a readily usable quantitative framework for organizing diverse nutrient data sets.  The essential parts of simple budget include inputs (point, diffuse, atmosphereic) and exports (denitrification, burial in sediment, oceanic exchange, fisheries harvest).  Additional elements that we should think about include nitrogen storages in the sediments, eroding marshes (a form of inputs) and nitrogen fixation (another potential input).


Typical anthropogenic nitrogen loads to watersheds along the East coast of the USA vary from about 2-10 gN/m2/yr.  However, many estuaries receive nitrogen loads well above this level, in part because the ratio of basin area to estuarine area are generally large (>5:1) and because direct inputs of N from point sources can be large. On a global sampling of estuaries there are actually a couple of orders of magnitude difference among the lowest and highest loaded estuaries. However, loading rates are not sufficient by themselves to indicate nutrient enrichment effects!  For example, the Baltic Sea has 1/10 the load of the Potomac River but exhibits some serious eutrophication

features, probably because the Baltic is so poorly fluched.  In a similar fashion the Patuxent River and Narragansett Bay have similar loading rates of N but eutrophication indices are more severe in the Patuxent, again because of longer residence times.  Lastly, loads may differ from year to year.  Interannual changes are on the order of a little over 2 for N and 3 for P and these interannual changes are large compared to the expected size of many, but not all, proposed management actions.



In the coastal bays there is a large range in loading rates.  Areal loading rates, based on the earlier work of Boynton et al (1993)for the coastal bays are as follows:


g N/m2/y

Assawoman Bay


Isle of Wight


St. Martin Rive


Turville Creek 


Sinepuxent Bay 


Chincoteague Bay


Newport Bay



Simple scatter plots of TN loading rates versus water column chlorophyll-a concentrations exhibit a positive relationship, as expected, suggesting that the relative magnitude of these loading rates are reasonably correct.


Internal losses of N and P have not been well measured for the coastal bays.  However, it is not likely that either long-term burial or extraction of N and P in the recreational or commercial fisheries are large N and P loss terms.  Little is known about the net nutrient flux at the ocean inlets and relaible estimates would require a hydrodynamic model and a great deal of nutrient concentration data.


Net Anthropogenic Nitrogen Inputs to Coastal Bay Watersheds…and Nitrogen Discharges from the Watersheds.         Tom Jordan

Anthropogenic nitrogen exchanges to the watershed occur via agriculture (fertilizer input, nitrogen fixation by crops), via atmosphere (nitrate deposition, ammonia volatilization and deposition); and via trade (food inputs/export, feed input).  Some of the net input may be discharged to the river.


Atmospheric nitrogen deposition was calculated from the National Atmospheric Deposition Program date for dry deposition.  Wet deposition is estimated as equal to dry deposition.  Ammonium deposition may be offset by ammonia volatilization from livestock waste and fertilizer. Nitrogen input includes wet plus dry atmospheric deposition of nitrate at a rate of 6.8 kg N ha-1 yr-1.


Fertilizer nitrogen is estimated by sales in the county and recommended application rates since actual applications are usually not reported.  County data is then prorated by land use type in the watershed.  Agriculture census data are used to estimate food inputs and exports, feed inputs and nitrogen fixation by crops.  Census data are reported by counties (1987 census) and include harvests, livestock populations and amount of agriculture land.  Data censored from county reports are estimated from state-wide data distributed across the counties according to land use proportions.  Nitrogen fixation is estimated from crop types and harvests.  Net input to croplands includes atmospheric deposition of nitrate and

ammonium and that assumes 50% of livestock waste N is volatilized as ammonia and the remaining livestock waste is applied to croplands.


Nitrogen from agriculture sources may take awhile to get to bays if in



Exchanges of food and feed N are calculated as the difference between

consumption and production within the county.


Half of the N excreted by chickens can be lost via volatization to N gas.  The remaining N in livestock waste is spread on croplands within the county of origin.  Food input is calculated as number of people X 6.3 kg N yr-1.


We estimated the discharge of water, total nitrogen, and nitrate from the watershed based on the percentages of cropland and developed land in the watershed and linear regression models that relate these land use proportions to watershed discharges.  The regressions were developed from measurements of discharge from 26 Coastal Plain watersheds on the western shore in Maryland.


In summary the biggest input of nitrogen to the watershed is chicken feed and the biggest output of nitrogen from the watershed is ammonia gas (maybe 60-70% of total net [after volatilization input]). 


The model-predicted watershed discharge of nitrogen is only about 12-16% of total net input anthropogenic N input to the watershed.  In other words, most of the N input to the watershed (about 80%) does not seem to be discharged in water flows to the Coastal Bays.  The net input that is not discharged may be denitrified or stored in groundwater, biomass, or soils.  Total load from the watershed to the bays is 494,973.5 kg N/year.



Total inputs

(kg N ha-1 total area yr-1)

Total N Discharge

(kg N ha-1 total area yr-1)

Total N Load

(kg N/yr)

Little Assawoman








Isle of Wight Bay




Sinepuxent Bay




Newport Bay




Chincoteague Bay







TMDL assessments of loads and Point source loads.         Jim George

Total nitrogen loads for the northern coastal bays were estimated to be as follows (based on 1998 Point source loads and 1997 land use).


Total Nitrogen Load




Isle of Wight and SMR


    St. Martin TOTAL


      >Shingle Landing Prong


      > Bishopville Prong


Herring Creek


Turville Creek


Newport Bay TOTAL


  > Ayer Creek


  > Newport Creek



Average annual loads were compared using rates from a 1993 UMCES report, Chesapeake Bay Program segment 430 and from a DNR/USGS study for the Pocomoke.  The loading rates of the DNR Pocomoke study were the highest of the three.  MDE used the three different loading rates in separate runs of the Northern Coastal Bay water quality model.  Modeling results associaed with the UMCES loading rates best replicated the observed water quality data.  This was used to justify using the UMCES loading rates modified slightly as discussed below.


Stormwater runoff loads from the higher density urban landuse on Fenwick Island was believed to be higher than the 4.4 lbs/acre suggested in the 1993 UMCES report.  This was based on data collected by the NPDES stormwater permitting program from Baltimore City and Baltimore County.  Consequently, the urbanized parts of Fenwick Island were assigned a 7.56 lb/acre loading rate.  This decision was indirectly substantiated during the Newport Bay model development process. Calibration of the Newport Bay water quality model was only possible after increasing the loading rate for the City of Berlin from 4.4 lbs/acre to 7.56 lbs/acre.




Northern Coastal Bays:  Nutrient Load Monitoring.          John McCoy

On the eastern shore of Maryland, ditches are a major feature.  Ditches circumvent the natural groundwater paths.  And, ditches significantly extend the number of miles of streams and increase nutrient flows. For example, in a Chesapeake bay watershed the streams were documented at 82 km but if the ditches were counted it would add another 195 km.


WRAS; yields and instantaneous loads (works for nitrate.. spring recharge period have highinputs.  Results from synoptic surveys show that baseflow can be indicative of average annual 10-20 kg/ha/yr in 1999 and exports of 60-70 kg/ha/yr in 2001.


The ratio estimator was used for Birch Branch to calculate actual loads (~4,000 acres watershed).  In  2003 the load was 38,105 and in 2004 it was 40,931 kg.  Nitrate/ nitrite accounted for almost 50% of the load (this gave yields of 20.57 lb TN/acre in 2003 and TN; 2004:


A long term record (ten years) of inputs to the Pocomoke River show loads to be highly variable (5-50 lb/acre) primarily due to water flow and stream management activities (ditch management)



A source apportionment of atmospheric deposition.            John Sherwell

Shallow groundwater on the Delmarva are among the highest in the nation.


A model named CALPUFF was used to track deposition from ~1000 sources.  The model allocates loads to different sources of atmospheric deposition to the coastal bays.  This model is good at analyzing “what if” scenarios for individual actions, it offers an ability for directed negotiation.  The next iteration will have “real” ammonia


Total atmospheric load of nutrient N to the Maryland Coastal Bays is about 320,000 kg-N/yr. A total N-deposition flux of about 10-11 kg-N/ha-yr [Boynton et al & Meyers et al].  This may comprise as much as 55% of the total anthropogenic input. [Boynton et al)

There is a large variability from year to year: Assateague data was lower than the site at Wye in 2001 AND 2002.  Dry deposition is estimated from CASTNet sites.  Also need measurements of organic nitrogen which is largely biogenic and poorly quantified. Nitrate inputs were between 1.7 – 2.9 kgN/ha/yr.  Modeled ammonium deposition (does not include ammonia measurement, hence LOW bias)


Results show that the EGU sector is the largest contributor (40.6%), followed by mobile sources (26.8%), then local area sources (24.5%) and finally industry (8.1%). Area sources include small stacs, jet skis, agriculture equipment, lawn mowers, chain saws etc..  Mobile sources include anything that drives on their own.  And industry category includes all tall stacs (Clean Air Act permits?).


Regional sources account for 58.4% of the total atmospheric inputs (MD 17.4%, VA 15.3%, Delaware 14.6% and PA 11.1%).  The top ten individual contributors are Sussex, Sussex, Grid?, Morgantown, Gen_JM_Gavin, Baltimore Gas and Electric, VA Power at Dutch Gap, Potomac Electric, Belew Creek, Mt Storm.


Wet (45.7% of load) = dry (54.3%) deposition is not a good assumption.


“Exhaust from two hours of jet skiing is equivalent to the smog-forming emission from a 1998 passenger car operated for about 130,000 miles.  About 30% of a jet ski’s fuel escapes into the air and water.”  (California Air Resources Board).



Transport of Nutrients from GW to the Coastal Bays.  Bob Shedlock

Based on ground-water flow-net analysis, recharge areas that discharge ground water directly to the Coastal Bays are small.


An interpreted salinity section across Chincoteague Bay shows a freshwater wedge extending beneath the bay from the western shore, and dissipating within one-half mile from the shoreline.  The section indicates brackish water beneath the bay bed near Assateague Island.


Nitrate in surface-water baseflow has been shown to be correlated to the percentages of row crop and deciduous forest in the watershed (Dillow and others, 2002).


Loads were calculated for Birch Branch and Bassett Creek during CY2000 and CY2003, respectively.  Using the USGS ESTIMATOR model, mean annual TN loads were 58 kg/day in Birch Branch, and 19 kg/day in Bassett Creek.  Mean annual TP loads were 3.6 kg/day and 2.2 kg/day at Birch Branch and Bassett Creek, respectively. Instantaneous TN loads followed streamflow levels in both systems, ranging from 6 to > 1000 kg/day at Birch Branch, and from 4 to >400 kg/day at Bassett Creek and, at baseflow, were higher in Birch Branch.  Baseflow was also higher in Birch Branch (generally >10 cfs).


Nutrient yields for Bassett creek during CY2003 were 12.4 lbs/acre/yr for NO2+NO3, 20 lbs/acre/yr for TN, and 2.3 lbs/acre/yr for TP.  Nitrate/nitrite were pretty much equally attributed to baseflow and quickflow, while 60% of TN yields and 90% of TP yields were from quickflow.


Monitoring of stream water quality below head-of-tide, including in small tidal tributaries, needs to be carried out in order to more accurately estimate the true nutrient loads being contributed to the coastal bays by streamflow.




Shoreline Erosion inputs.       Darlene Wells

The Maryland coastal bays shoreline is 452 miles long.  Based on changes in shoreline position in the 27 years between 1962 and 1989, the shoreline, on average, is retreating about 1 foot every 5 years (-0.19 ft/yr or -0.6 m/yr).  However, erosion tends to be highly variable spatially and episodic in its occurrence.  Over the entire period of study, 1850-1989, the coastal bays lost 3,758 acres to erosion.  This was offset by a “gain” of 1017 acres along the eastern shore of Sinepuxent Bay, as Assateague Island migrated landward.  That translates into an average annual loss of 27 acres throughout the coastal bays, again offset by a 7 acre gain along eastern Sinepuxent Bay.


About 10% of the shoreline baywide is currently protected by erosion control structures, though that number varies considerably depending on the bay (L. Hennesee MGS).  The northern coastal bays have a larger percentage of protected shoreline - anywhere from 20-40% more than do the three southern bays. 


MGS conducted a study during which they calculated the nutrient (total carbon, nitrogen and phosphorus) contributions to the northern and middle bays from shore erosion  (Wells et al, 2002, 2003). The southern half of Maryland's portion of Chincoteague (south of Tingles Island) was not included.  For the study, MGS measures bank heights and collected sediment cores along various natural shorelines characterized by higher erosion rates (primarily marsh shorelines).  The sediment cores were analyzed for bulk density properties, and nutrients (total carbon, nitrogen, and phosphorus).   Bank height and sediment data were used with shoreline change data to calculate annual rates of nutrient input from shore erosion.   Results from the study indicated that shore erosion in the northern and middle coastal bays contributes approximately 4% of the total nitrogen entering those bays.  Shore erosion contributes 5% and 9% of the total Phosphorus entering northern and middle bays respectively.  N and P loads from other sources were based on Boynton (1993).

Nutrient loads


N. Coastal Bays

Mid. Coastal Bays

Southern Coastal Bays ???

1989 shoreline length (m)




Total Sediments (Kg/yr)




Total Carbon (Kg/yr)




Total Nitrogen (Kg/yr)




Total Phosphorus (Kg/yr)





C, N and P contributions for the southern coastal bays (southern half of the ChincoteagueBay) were estimated using the following values from the middle coastal bays study (Wells and others, 2003). 

•Average bulk density of shoreline sediment            0.79 g/cm3  

•Average phosphorus for marsh sediments              0.034 %. 

•Average nitrogen for marsh sediments                    0.25 %

•Average carbon for marsh sediments                     4.26 %


Some problems (caveats) with N and P loadings calculated for southern Chincoteague include:

1) Calculation assumed all shoreline is eroding; for northern and middle bays, only eroding shorelines were included in calculations

2) Geomorphology and underlying geology in area are different

3) N and P loadings calculated using different method than northern and middle coastal bays.


Bulk component properties for each station sampled for sediment nutrient studies for the northern and middle coastal bays were plotted against northing UTM coordinates.  Plots suggest trend in decreasing carbon contribution in a southerly direction.  However, phosphorus seems to increase going south.



Monitoring of sediment oxygen and nutrient exchanges in Maryland’s Coastal Bays in Support of TMDL Development. Walter Boynton

A total of 21 TMDL stations were monitored during the June, July and August of 2003 (5 Assawoman, 3 St. Martin River, 6 Isle of Wight Bay, 1 Sinepuxent Bay, 2 Newport Bay, 1 Trappe Creek, 1 Marshall Creek and 2 upper Chincoteague Bay).  Cores were incubated in the dark at ambient water temperatures.


Sediment Oxygen Consumption in the tributaries averaged ~3 g O2/m2/d (ranging from 1.2 – 4.3) and was between 1-2 g O2/ m2/d in the open bays.  All 21 sites gave an average of 1.5 g O2/m2/d.  In comparison the upper anacostia on the order of 4 (highest).


Sediment ammonia flux results show that some areas are a sink but most fall between 0 and 100 μmoles N/m2/h.  With the exception of a deeper site in Manklin Creek that averaged almost 700 μmoles N/m2/h.  In summary the sediments are not releasing large amounts of ammonium in general (relationship SOC : ammonium flux 13:1 anatomic) something else is happening. Different than the deeper Chesapeake which is highly hypoxic and out of light.  Some NH4 may be taken up by autotrouphs, nitrification with diffusion into the sediments.  This type of environment tends to promote recycling.


Phosphorus:  very modest releases of phosphate on the order of 5 – 10 μmoles P/m2/h.  sediments largely oxidized.  Iron present, oxidized iron binding to P.   Such levels are not going to grow a good bit of phytoplankton (generally need to get near 40-50 μmoles P/m2/h).



Wrap up


1.          Create table of different estimates with consistent units. ( DNR with help from authors)

2.          Nitrogen verses Phosphorus limitation.  Use Fischer model.  (DNR lead)

3.          Determine nutrient burial and denitrification estimates using MGS data and rates from Boynton and Cornwell.  (MGS and UMCES)

4.          Compare model loadings for Bassett Creek and Birch Branch sub-watersheds to measured loadings by DNR and USGS/NPS.   (DNR)

5.          Layer GIS problem areas (Niles) and poultry houses; row crops and communities on septic could be digitized and related to groundwater NO3. (DNR, WC?)

6.          Cost efficiency of chicken waste transport is 12 miles. Put 12 mile radius around poultry houses in GIS.  Check if old digitization of poultry houses accurate (not all houses have chickens – newer houses should have manure sheds nearby) ??