Explanation of Nitrogen Loading Assumptions and Coefficients
by Joe Costa
Please report errors or send comments about this page to Joe Costa.
Nitrogen loading spreadsheets and models are now routinely constructed and evaluated by managers, consultants, regulators and scientists to approximate nitrogen inputs to coastal waters and groundwater from various types of land use. These loading spreadsheets are based on assumption of average or typical loading for different types of land use or nitrogen source because it is impossible to quantify actual loadings from every diffuse source of pollution. Sometimes these loading estimates are customized using parcel specific information like impervious area, lawn area, or municipal water use records.
Our Nitrogen Management page has a summary table of nitrogen rates for the principal nitrogen sources in most watershed. On this page we explain in more detail some of the assumptions used in the various loading models, and provides guidance on when some of the assumptions may be changed. We show some approximations for some loading coefficients so that municipal can quickly evaluate the correctness of the data submitted. We also provide some guidance to municipal boards and regulators on data and loading analysis submission requirements to ensure that nitrogen loading calculations are transparent, understandable, and can be validated with the information provided. We also provide guidance on how to review these data sets and to catch potential errors or oversights. For non-ArcMap users, the MassGIS Online Mapping System (Oliver) may be useful for determining the areas of certain types of land use. On various pages we also provide Google Earth files of subwatershed data. On this and other pages we provide loading rates in both English (pounds per acre) and metric (kilograms per hectare) units.
In the 1980s and 1990s, scientists, planning groups, and regulatory agencies developed various nitrogen loading models and spreadsheets to evaluate nitrogen inputs to groundwater and coastal waters from existing development, and potential inputs from proposed new development. In the 1980s, the Cape Cod Commission Planning and Economic Development Commission evaluated nitrogen loading to groundwater, first for planning purposes, then for regulatory purposes when they became the Cape Cod Commission. In 1991, the Buzzards Bay NEP developed several loading spreadsheets to estimate nitrogen loading to coastal waters based on either town parcel data or coarse evaluations of land use, like MassGIS land use based on the interpretation of aerial photographs. The Buzzards Bay NEP used loading coefficients from the Cape Cod Commission and from other sources. A list of embayments and their degree of overloading was included in the 1991 Buzzards Bay CCMP.
By 1994, the Cape Cod Commission was requiring nitrogen loading evaluations for Development of Regional Impact using prescribed loading assumptions. A special nuance of the CCC approach was that it was based on development not exceeding a 5 ppm groundwater concentration at the property boundary. This meant regulatory calculations were based on an estimated discharge of nitrogen inputs to groundwater, divided by assumed annual rainfall recharge over the entire lot. (In some cases this led to “gaming” of the method to increase impervious area size to increase recharge and thus lower boundary nitrogen concentrations). On the other hand, the Buzzards Bay Project recognized that a 5 ppm groundwater standard was woefully inadequate to protect impaired waters, and noted that groundwater target must be 1 ppm or less to protect most coastal embayments.
By the late 1990s, new data and information on both loading and impacts were becoming available. In the 1980s and 1990s, it was presumed that nitrate traveled great distances without loss (unattenuated). While this is still generally believed true, it is now recognized that as groundwater enters fringing wetlands and ponds in the upper watershed, or even as it passes through marine sediments, some of that nitrogen is removed by plants and microbes (attenuated). In the absence of adequate studies, the BBNEP first used a zero loading coefficient in 1990, a 33% attenuation loss by the mid 1990s, and a 50% attenuation loss by the 2000’s to be consistent with the Massachusetts Estuaries Project studies.
In 2001, DEP began implementing the Massachusetts Estuaries Project with the University of Massachusetts at Dartmouth School of Marine Science and Technology (SMAST) to establish nitrogen TMDLs in 89 Massachusetts coastal embayments. When the program commenced, the SMAST reviewed the BBP, CCC, and other efforts to estimate nitrogen loading. Based on their review, SMAST updated many of the loading coefficients, and decided to apply the standards to municipal parcel information. In the sections below we describe the MEP assumptions for each category and variations of methodology that have been applied by the MEP and others.
Model Robustness and Linkage to Standards
Nitrogen models are often robust with respect to management action, even if their loading assumptions differ greatly. To illustrate this point, in the 1990s the BBNEP noted that most eelgrass disappeared in a particular embayment in the mid 1960s when there were roughly 1200 homes built in the watershed. We suggested a nitrogen loading target limit based on our nitrogen loading model at the time, and that the town should work to reduce inputs (principally septic systems) according to this standard. Subsequently, a new scientific study in the watershed concluded that perhaps up to 50% of the nitrogen assumed to be reaching the bay were removed via various mechanisms. This led to the comment by a resident that nitrogen reductions were therefore not required. Such thinking is flawed, because the new analysis did not restore the system and cause the recovery of eelgrass. Rather, it illustrates the point that recommended nitrogen thresholds and ecosystem response models are fundamentally linked to whatever nitrogen loading model on which they are based. In this example, the threshold, using the new model would be half that previously recommended, but it would still be equal to the nitrogen loading of 1200 residential units in the watershed. Thus, to determine whether recommended TMDLs are being met, the applicant must use the MEP loading model if that is the basis of the TMDL study. Moreover, if appreciable changes in the loading assumptions are made (septic loadings, unaccounted for lag times, changes in watershed attenuation coefficients, etc.), remodeling of the ecosystem response may be necessary to calculate a new TMDL threshold limit.
Watershed nitrogen loading models are just one required element for establishing estuary management limits for nitrogen, referred to under federal water quality regulations as Total Maximum Daily Loads (TMDLs). In Massachusetts, these watershed nitrogen loading models are linked to an estuary water circulation model to help predict average annual total nitrogen concentrations (or perhaps summertime conditions) at a sentinel stations. If the protection of eelgrass is a management priority in an estuary, a target nitrogen concentration at the sentinel station would be selected to ensure much of the estuary (on average) would be at or below 0.4 ppm total nitrogen. If eelgrass were not a management priority, and the integrity of fish, shellfish, and other animal communities were the management priority instead, a nitrogen concentration at the sentinel station would be selected to ensure that much of the estuary would achieve 0.5 ppm total nitrogen or better. The actual total nitrogen target concentrations selected may also depend on other factors, and issues related to setting the TMDL are discussed elsewhere on this website.
Ensure Septic Loadings are Correct
MEP Loading Analyses and Review
The BBNEP has long recommended parcel level land use analyses as the basis of nitrogen loading analyses, and this is the basic approach of the MEP model. Many variations exist on this approach. For example, the MEP expends a considerable effort obtaining parcel data, but most of this parcel data is reaggregated by assessors land use codes for the estimates of actual loading. This creates a more manageable size spreadsheet for large watersheds, but the loading spreadsheet looses the ability to easily override assumptions for individual parcels.
In the sections below we identify loading assumptions used in MEP and other loading models. While some of these assumptions could be scientifically challenged, it is important to recognize that many of these nitrogen sources are relatively unimportant in most watersheds. In nearly every watershed, however, the most important calculation is the number of septic systems in a watershed. This means that municipal officials should focus on reviewing GIS data and maps showing existing conditions, particularly whether sewered and non-sewered parcels were correctly identified, and whether the number of existing residential units (septic systems, lawns, roofs) were correctly enumerated. Some common pitfalls in calculating septic loads are illustrated in the two figures to the right. There are also potential confounding factors associated with water data, but these limitations can be addressed, and will be discussed elsewhere.
Key Types of Land Use and Nitrogen Sources
Note: all the applications rates defined below may have an addition loss (attenuation) term applied to depending on how far the source is from the estuary or receiving waters. The attenuation coefficient is often set at 50%. Alternatively, MEP and other may use actual flows and concentrations of stream flow into an estuary to calculate loadings for all land use above that sampling point, and only use parcel level loading analysis for areas close to an estuary.
Wastewater discharges, whether via municipal wastewater facilities to surfaces waters or groundwater, or individual onsite residential septic systems are individually and cumulatively, generally the largest source of nitrogen to coastal embayments, typically accounting for 60% to 85% of nitrogen loads. It is for this reason that the assumptions relating to wastewater discharges are the most critical to any watershed nitrogen loading analysis, and for deciding the best nitrogen management strategy.
Wastewater Treatment Facilities
All municipal wastewater treatment facilities that discharge to surface waters must monitor nitrogen concentrations. This data is available the US EPA PCS system discharge permit data website (click on permit number to see data) as well as the EPA NPDES reports website. To calculate existing discharges, actual flows and concentrations should be used, whereas for build-out scenarios, permit discharge volume limits should be used, as eventually most towns will eventually reach those limits.
In terms of evaluating the validity of ecosystem response models, a critical nitrogen loading evaluation decision is what period of discharge to use: summertime loading or year round average loading rates? This decision is important because, while discharge volumes of wastewater plants may increase somewhat during the summer, nitrogen removal capacity often improves more dramatically because higher effluent temperature improves biological removal processes.
In the case of state issued groundwater discharge permits of wastewater (automatically required for discharge permits of 10,000 gpd), the Commonwealth of Massachusetts is increasingly automatically requiring discharge limits for nitrogen, limiting discharges to 10 ppm or less to nitrogen impaired coastal waters, and also requiring effluent monitoring for nitrogen. Because there is a lag between discharges to groundwater and when that groundwater reaches coastal embayments (and the rate of groundwater varies between wet and drought years), it is always best to use annual average loadings based on each sampling dates flow and concentration.
Estimates of onsite wastewater disposal systems, commonly referred to as septic systems, are the most important element of any nitrogen loading model because septic systems the typically account for 50% to 90 of the nitrogen loading in most watersheds. After review various literature sources, the in the 1990s the Buzzards Bay NEP adopted a per capita loading rate of 5.95 pounds (2.7 kg) per year reaching groundwater (actually an adult loading rate; explained in this 1999 report Managing anthropogenic nitrogen inputs to coastal embayments (1.2 MB pdf). The Buzzards Bay NEP used US Census and town population statistics, including estimates of summer occupancy in its watershed loading estimates.
The Buzzards Bay NEP adopted a per capita approach because we long recognized that septic system effluent nitrogen concentration was highly variable. In households with high or wasteful water use and older home that lack low flow devices may have wastewater concentrations less than 30 ppm. Households with more water saving devices and frugal water use may have effluent concentrations of 60 ppm or higher.
The Buzzards Bay NEP adopted the 5.95 pounds per capita nitrogen loading to groundwater because it was also consistent with the Cape Cod Commission methodology adopted for regulatory purposes in the 1980s. This agency used effluent concentrations of 35 ppm N and a system flows of 55 gallons per day per capita flow in its assessments to protect drinking water supplies. The 35 ppm concentration adopted by the Cape Cod Commission has since been criticized as too low based on the literature, but their selected 55 gpd discharge rates were are about 25% higher than the typical 44 gpd septic system discharge reported by EPA at the time. In addition, in its regulatory review, the Cape Cod Commission used a much high worse case occupancy rates (intermediate between typical and maximum occupancy of 2 persons per bedroom), adding a considerable margin of safety for their evaluations.
In 2001, the MEP adopted a very different approach to estimate nitrogen loading because they desired site-specific estimates for small Cape Cod subwatersheds where their initial studies focused. Because Cape Cod had both high seasonal and weekend population fluctuations, the MEP used municipal water department records of town water use for estimating septic system loadings. When town water use is not available (e.g., residences with private wells), the MEP uses an assumed average daily flow single family residential in the watershed or town, or nearby town. In such cases, the MEP calculated average water use for single family residences then multiplied this volume times an assumed septic system discharge groundwater of 26.25 ppm. This value was based on Costa et al. 2002 and other studies that documented leach field nitrogen losses from conventional septic systems.
It is worth noting that in the MEP reports and spreadsheets, septic loading is calculated by either 1) multiplying total residential water use by 90%, then by 26.25 ppm, or more commonly by 2) multiplying total water use times 90% of 26.25, which is 23.625. In most TMDL reports, the authors write, “As a result, MEP staff has derived a combined term for an effective N Loading Coefficient (consumptive use multiplied by N concentration) of 23.63, to convert water (per volume) to nitrogen load (N mass).” This means they took the total billed water use and multiplied it times 23.625 ppm to calculate total septic loading.
There are some highly nuanced and challenging issues related to interpreting water use records, the biggest challenge of which is account for sub-watershed variability in lawn irrigation water use, and these are discussed on our occupancy and residential water use page. The MEP generally presumes that 90% of water use discharged to septic systems, irrespective of watershed. An interesting outcome of the MEP water use approach is that it implies a variable per capita nitrogen loading rate based on seasonally adjusted census populations, if the census data is taken at face value. In the Eel Pond/Back River watershed report, MEP calculated a 182 gpd per Single Family Residence water use and a discharge concentration of 26.25 ppm (x 0.90). This results in a annual loading rate of 13.0 pounds per year per single family unit. With an occupancy rate of 2.35 person per occupied unit, this implies a per capita rate of 5.53 pounds per person. The actual per capita value implied by the MEP analysis is actually higher, because water use is integrated annually for both occupied, unoccupied and seasonally occupied units. The table below shows actual pounds per residential unit used in various MEP studies, with the implied per capita rate based on average occupancy rate of occupied units. Actual town-wide average annual occupancy rates (including weighting for summertime population increases) is around 2.49 persons per all units in the Town of Bourne (view our demographics and water use webpage). US census data exists at the block level, so US census data could actually be calculated for MEP subwatersheds.
A starting point, or point of comparison for many MEP water use calculations is the each person discharges on average 55 gallons per day to a septic system Actual water use varies from town to town, and municipal estimated per capita water use is more typically in the range of 55-75 gpd per capita (see water use data reported to DEP). A 55 gpd volume, times 26.25 ppm TN is equal to 2.0 kg (4.41 pounds) per year.
|MEP calculated average water used for Single Family residences served by a public water department.|
|MEP Study||MEP observed Water Use (gpd)||MEP est. non-waste water use||MEP est. septic flow gpd||conc (ppm)||MEP lbs/unit/yr||MEP/USCen estimated
|implied pounds per person|
* This was based on a town-wide average in the draft MEP report. In 2010, the BBNEP obtained water use data from the Town of Wareham, and actual average water use in the watershed is considerably less. The BBNEP provided this data to the MEP, and the values used in the final TMDL report are expected to be lower.
Values Corrected. Based on interviews on Cape Cod, the MEP concluded homeowners only fertilized an average of 1.44 times per year and at rates less than recommended. They concluded homeowners 1.08 pounds of nitrogen per 1000 sq. ft. or 5.4 pounds annually per typical 5000 sq. ft. lawn. With a 20% leaching rate based on various studies, this results in an overall annual loading rate of 1.08 pounds per lawn. This equals 9.41 pounds per acre. Text and loading spreadsheets by the MEP note that they assume only 1/2 of all lawns are fertilized. Previously we incorrectly stated that they the MEP halved the 1.08 pounds per 1000 sq. ft. value in their loading estimates, but this is not the case based on a review of recent data disks prepared by the MEP for their TMDL reports, and the 50% fertilization rate may already be built into their loading assumptions.
Parks and Schools
The MEP assumes 3.0 pounds of nitrogen per 1000 sq. ft. with a 20% leaching rate. This equals an annual loading rate of his equals 13.50 pounds per acre (15.13 kg/ha).
The MEP dedicates a considerable effort in characterizing golf course loadings by digitizing the acreages of all the subareas of golf courses, including tees, greens, fairways, and rough, and applying different nitrogen loading coefficients to each. When course-specific fertilizer use is not available, they employ average loading rates based on a survey of fertilizer use of Cape Cod golf courses, and these coefficients have changed among their studies. Generally course-specific fertilizer use data is not available, so these latter values will be used.
Below is a summary of the loading calculations for the Brookside Golf Course used for the in their Eel Pond/Back River MEP report. This estimate is based on the average fertilizer applications for the six Cape Cod golf courses for each turf type, multiplied times the area of each turf type determined from aerial photographs (map to the right). Although the percentage of rough and fairway may change considerably from course to course, the average per acre loading for this entire course shown (23.8 lbs/acre). Because golf courses are generally a small percentage of most watershed nitrogen loadings, an average rate of 24 lbs per acre to the entire course area will likely be a reasonable approximation of golf course N loadings in most watersheds.
|golf area||sq ft||acres||% of
|lbs applied||lbs leached
|lbs per acre leached||23.83|
The acreage of the Brookside Golf Course based on MEP’s GIS coverage is 70.5 acres (a small portion of the course is outside the watershed and not in the table above) as compared to 85.2 acres calculated from MassGIS 2005 land use data (map above). Golf areas in the MassGIS database will be larger than direct measurements of sub areas because MassGIS land use includes all areas of the property including clubhouses and property to road boundaries.
Updated. Up until 2010, the MEP used in their TMDL watershed nitrogen analyses an annual nitrogen application rate of 31 pounds per acre with 66% leaching, for a net loading rate of MEP uses a loading rate of 20.46 lbs/acre. As of 2011, draft reports now use 6.2 lbs per acre based on a 2005 study by Howes and Demoranville. The purpose of this study was to evaluate Phosphorus loading from surface water discharges from bogs, and groundwater nitrogen discharges were not evaluated. Thus the 6.2 lbs per acre may underestimate actual watershed nitrogen loads from cranberry bogs.
In some studies, MEP has used the acreage of Chapter 61 lands at the assessor’s office, and multiplied this acreage by 85% to estimate bog acreage. As shown in the map to the right, this resulted in an overestimate in cranberry bog production area in the Pocasset Back River Eel Pond watershed where MEP estimated bog production area as 55.3 acres, whereas actual bog acreage was 27.1 acres. For watersheds where agriculture is a large source of nitrogen, calculation of actual production area is essential. At the same time, agricultural staging areas, bog berms, tail water ponds, and sand mining areas should not be treated as having the same loading rates as natural landscapes.
MEP assumes 30.49 lb/ac typical application rate and 30% leaching. This results in a net loading rate of 9.10 lbs per acre (=10.20 kg/ha). Agriculture in SE Massachusetts typically include corn, strawberries, and mixed vegetables. This loading rate should be made more watershed specific. For example, a 30 lb per acre application rate for corn may be unreasonably low, and is likely triple this rate. If agriculture dominates a particular watershed’s landscape, the USDA Agricultural Chemical Usage statistics website will prove valuable. For example, in the five states of Minnesota, New York, Oregon, Washington, and Wisconsin, in 2006 the average sweet corn nitrogen application rate was 93 pounds per acre. For farm-specific phosphorus evaluations, the USDA Phosphorus Loading Tool will be useful.
MEP assumes 4.46 lb/ac nitrogen application rate and 100% leaching (=5 kg/ha).
For impervious surfaces, MEP uses 90% recharge. While this value might be realistic on Cape Cod in situations where rainwater is discharged to dry wells or infiltrating catch basins, it may be far to high for watersheds with C type soils. A more practical interpretation of this loading coefficient is that 90% of impervious surface runoff reaches groundwater or surface water depending on watershed hydrology.
The MEP sometimes may use town specific annual rainfall, which is important because there is a strong precipitation gradient in Southeastern Massachusetts, declining from the Rhode Island Border to Outer Cape Cod. For example, New Bedford has annual precipitation of 50.8 inches, East Wareham has 48.8 inches, and Hyannis has an annual precipitation of 43.0 inches (see NWS 30 year normals for MA, a 1 MB pdf file). Values of annual precipitation used in MEP reports have been 44.4 inches for all Cape Cod studies. Adding several inches to annual precipitation to impervious surfaces, ponds, and natural surfaces will increase total watershed loadings by several percent using the MEP nitrogen loading model.
Roads, Driveways, and Parking Lots
For roads, driveways, and parking lots, the MEP assumes 90% of watershed annual precipitation recharge rate (or discharge to surface waters) at a concentration 1.5 ppm, for an effective loading rate of 13.50 lbs per acre impervious. Because town parcel road coverage is for the road layout, not the paved surface (typically 75% of the layout), some adjustments must be made for the impervious area depending on whether paved sidewalks or grass strips adjoin the road.
MEP spreadsheets often have an assumed 1500 sq. ft of driveway area for all properties, but in some watersheds, other values may be used. In their reports the MEP does not explicitly address other small impervious surfaces that may occur on properties (walkways, patios, porches), but these are generally much smaller by comparison to driveways, and these could be considered incorporated within a driveway area calculation. Overall these assumptions appear to be reasonable average estimates for single family residential properties, although some watersheds may have different values. In some reports these values are sometimes applied to larger commercial and public properties that may far larger roof, parking, and driveway areas. On some MEP loading spreadsheets, however, it appears that MEP does not actually always use the driveway area value in their watershed nitrogen loading calculations, so there may be some variability in the actual application of this parcel coverage.
The MEP assumes 90% of watershed annual precipitation recharge rate (or discharge to surface waters) from roofs at a concentration 0.75 ppm, for an effective loading rate of 6.76 lbs per acre impervious. MEP has often used an average residential structure footprint of 1500 square feet, based on a study of town of Falmouth structure information, for its calculation of roof runoff. However, if footprint of parcel structures is available in GIS data set, or in an assessor’s databases (rare), these impervious areas may be directly calculated in the MEP loading spreadsheets.
Saltwater (embayment surfaces)
For precipitation loading to saltwater surfaces, the MEP assumes 90% of annual precipitation (rather than an expected 100%; e.g. 40 inches for the Pocasset Eel Pond and other Cape Cod studies) times a presumed 1.09 ppm nitrogen concentration in precipitation. This equals a loading rate of 9.82 pounds per acre (=11.00 kg/ha).
Fresh Surface Waters (again updated 16 February 2011)
For precipitation loading to freshwater surfaces, the MEP assumes the same rate for nitrogen deposition to coastal embayment surfaces, namely 90% of annual precipitation (rather than an expected 100%) times a presumed 1.09 ppm nitrogen concentration in precipitation. This equals a loading rate of 9.82 pounds per acre (=11.00 kg/ha). The MEP treats bordering vegetated wetlands as “natural landscapes” in their analyses, except in the Slocums River analysis, where they included all vegetated wetlands in the open water category to match the observed high nitrogen loads documented in the River systems. This was an important deviation from their standard model because there are 639 acres of water body surface area in this watershed (2/3 of which are estuary) but 6,178 acres of wetlands plus open water. The adoption of this approach in the wetland dominated watershed increased unattenuated watershed loading by roughly 40%.
For natural surfaces, the MEP assumes 27.25 inches of rainfall recharge (roughly 60% of annual precipitation on Cape Cod) and a recharge nitrogen concentration of 0.072 ppm. This equals a loading rate of 0.45 pounds per acre (=0.5 kg/ha). The definition of natural surfaces includes undeveloped lands, and portions of developed parcels not included in calculations for driveways, roads, roof area, and lawn. Natural surfaces include both uplands and wetlands in the MEP analyses, except in the Slocums River TMDL study, where wetlands were lumped with water body surface area as described above.
Wastewater Treatment Facilities
All Buzzards Bay municipal wastewater treatment facilities must monitor nitrogen concentrations, and most must report loadings to the EPA as pounds per day. Because there may be seasonal variability of both flows and nitrogen concentration, particularly in the case of facilities with advanced nitrogen removal, the MEP weights monthly data to estimate annual loading. One of the biggest challenges in assessing annual loading is that municipalities monitor nitrogen only on a monthly basis, and data may be highly variable without distinct seasonal trends.
The MEP reports contain buildout analyses that are principally based on property ownership (public versus private) and assessors land use codes (those considered developable). In some MEP reports in the Buzzards Bay watershed (e.g. Little River report available at the Oceanscience.net website), they may not always account for certain types of protected open space coverages like Conservation Restrictions (CRs) and Agricultural Protection Restrictions (APRs), or DEP Wetland Conservancy Program coverages. In watersheds where these areas are extensive, this can result in overestimates of buildout potential. For these reasons, MEP report buildouts should be considered as rough thumbnails as to how intense future development may be. It is important to recognize the MEP buildout analyses are not used to establish the value of the TMDLs. They are meant to both illustrate how overloaded ecosystems may become without nitrogen management, and to point to areas of the watershed where future nitrogen management actions may be required.