Title: a Watershed-Scale Biogeochemical Loading Model for Nitrogen and Phosphorus

Basic Information

Title: A Watershed-Scale Biogeochemical Loading Model for Nitrogen and Phosphorus

Project Number: G-12

Start Date: 9/1/2000

End Date: 8/31/2003

Research Category: Water Quality

Focus Category: Hydrology, Models, Nutrients

Descriptors:

denitrification, ecosystems, hydrologic models, geographic information systems,

land-water interactions, land use, mathematical models, rainfall-runoff

processes, watershed management

Lead Institute: New York State Water Resources Institute

Principal Investigators: Robert W Howarth, Elizabeth W. Boyer, Dennis Swaney

Title: A Watershed-Scale Biogeochemical Loading Model for Nitrogen and Phosphorus

Principal: Robert W. Howarth, Dept. of Ecology and Evolutionary Biology, Cornell University, Ithaca, NY, 14850 and The Ecosystems Center, Marine Biological Lab, Woods Hole, MA 02543

Problem and Research Objectives

Two recent reports from the National Academy of Sciences have concluded that eutrophication is the biggest pollution problem in the coastal marine waters of the United States (NRC 1993, NRC 2000). Eutrophication lowers biotic diversity, leads to hypoxic and anoxic conditions, facilitates harmful algal blooms, causes dieback of seagrass beds, and can lead to changes in ecological food webs that lower fishery production (NRC 2000). Over 40% of the estuaries in the country are degraded from eutrophication, with the problem being particularly severe in the northeastern and mid-Atlantic regions (Bricker 1999). For most estuaries in these regions, eutrophication is caused primarily from over-enrichment with nitrogen; phosphorus is a secondary contributor (Howarth 1988; Nixon 1995; NRC 2000). Most of the nitrogen delivered to coastal waters in the US, including the northeastern and mid-Atlantic regions, comes from non-point sources in the watershed (Howarth et al. 1996). Agricultural sources are important in some watersheds, dominating the flux in the Mississippi River basin and contributing to the flux of some estuaries in the mid-Atlantic region, but atmospheric deposition of nitrogen from fossil-fuel combustion is an even greater source of nitrogen to estuaries for most of the mid-Atlantic region and for the northeastern US (Howarth et al. 1996; Smith et al. 1997; Jaworski et al. 1997; Goolsby et al. 1999; NRC 2000).

In regions subject to changes in land use and in atmospheric deposition of nitrogen, the processes that control nutrient loads to the coastal zone are complex. Variability of these hydrological and biogeochemical processes is increasing as weather and climate change. Understanding how these processes affect the magnitude and transformations of the nutrient loads is necessary in order to manage the environmental resources of the coastal zone. Further, it is important for those living in and managing coastal watersheds to understand the impacts of their activities and policies on these nutrient loads. A relatively

simple modeling tool that can estimate the impacts of various activities in the watersheds can greatly enhance, at low cost, our ability to manage these regions effectively and to communicate the effects of human activities and environmental processes on nutrient loads. The report of the National Academy of Science s Committee on Causes and Management of Coastal Eutrophication concluded that no model currently available to managers fulfills this need for estimating the controls on nitrogen loads (NRC 2000).

They noted in particular that most models used by watershed and estuarine managers fail to deal adequately with nitrogen deposition onto the landscape with subsequent export downstream, even though this is the number one input of nitrogen to many estuaries. The Committee further concluded that the development of such a model particularly one that deals with atmospheric deposition -- is one of the most pressing priorities for solving the problem of coastal eutrophication (NRC 2000). Our aim has been to develop such a model.

To mitigate the effects of human activities on the supply of nutrients to surface waters, managers are tasked with gaining an understanding of the landscape source areas delivering nutrients to receiving waters. We have developed an easy-to-use model for calculating loads of N and P to coastal watersheds, targeted toward management applications. The model describes transport of water, sediment and nutrients from the landscape to receiving waters. Our goal has been to create a model structure that will be used widely; thus we have developed the model in a commonplace platform: the Excel workbook. This version of the model, GWLFXL1.0, runs as a Visual Basic for Applications (VBA) program with an Excel interface.

Model Summary

In its current form, the model uses the event-based dynamics of a simple, lumped hydrologic model (Generalized Watershed Loading Function (GWLF) (Haith and Shoemaker, 1987) GWLF is a parsimonious, event-based model that has been used successfully to analyze the hydrology, sediment, and nutrient loads of several mixed watersheds in the United States, including the New York City reservoir system, the Hudson River (Howarth et al., 1991; Swaney et al., 1996), the Tar-Pamlico (Dodd and Tippett, 1994), and the Choptank River drainage of the Chesapeake Bay (Lee et al., 2000). We have added additional functionality to handle atmospheric deposition of nutrients, simple estimates of denitrification rate, and changes over time of the areas of different landuse/land cover categories. The original model used daily historic or synthetic temperature and precipitation data to simulate monthly discharge, sediment load, and nutrient transport. We have developed a separate stand-alone weather generation package (also Excel/VBA based) to allow the user to generate alternate climate scenarios in a format compatible with the model.

New Features

Model Input/Output After the “port” of GWLF code to Excel was achieved, several features of i/o were radically redesigned in the interest of flexibility:

Model simulation options are now controlled primarily from an Excel pulldown menu (GWLFXL) which appears when the workbook is loaded.
Model parameters can now be read either from existing GWLF input files (ie text files) or from parameter worksheets embedded within the workbook.
Model output is now organized into several output worksheets, depending upon the time scale desired (ie annual, monthly, or daily). Worksheets that group the output by land use category are also generated at the option of the user. An advantage of organizing model output by worksheet is the ready creation of graphics within Excel from the tabulated values, or further user-generated statistical analyses of model scenarios.

Model Calibration Mode. A major addition to the package is the model calibration mode which utilizes the Solver addin feature of Excel to obtain a least-squares fit of a selection of model parameters to monthly streamflow, sediment flux, or nitrogen flux data. Model parameters are selected and calibration datasets are entered in the calibrate worksheet. The desired calibration mode is chosen from the pulldown menu. Solver then drives the model, iteratively changing the selected parameters, until model best matches the data in a least-squares sense. Up to 5 parameters may be selected, though as of this writing, the procedure appears to work best with one or 2 parameters at a time.

Parameter Uncertainty Analysis. Another new mode of using the model is parameter uncertainty analysis, in which the effect of uncertainty about parameter values on model output is estimated quantitatively. The process occurs in 3 steps:

In the “stochastic” worksheet, the model parameters to be investigated are assigned probability distributions, together with estimates of their mean and variance, etc
The user chooses the number of replicate runs desired for the analysis, and then draws the corresponding parameter values from their individual distributions; this option is selected from the pulldown menu
The user runs the model in uncertainty mode, repeating the simulation for each realization of the parameter values, and the mean and standard deviation of the model outputs are stored in the “uncertainty” worksheet.

When the runs are complete, the user can plot the time series of means and confidence intervals for any model variable corresponding to the selection of parameters evaluated.

Project updates and website

The current version of the model and associated documentation and tutorials can be downloaded from the project website: Model updates, fixes, and future documentation will be made available here as well. While the VBA module containing the code is currently password protected to prevent tampering, the code is provided in Appendix 1. Interested researchers can obtain the password by contacting Dennis Swaney at .

Current and future research directions

Although the USGS/WRRI funded phase of the project has ended, we have obtained additional funding from an EPA star grant to pursue model development. We are currently engaged in adding more functionality to the model, aiming in particular at refining the descriptions of watershed biogeochemistry and hydrology, writing a model description for publication in a peer-reviewed journal, and beginning to evaluate the model against estimates of nitrogen load for 16 northeast US watersheds (Boyer et al., 2002). Links to further progress with the model development will be reported at the above website.

Publications and presentations associated with the project, 2002-2003

Boyer, E.W., C. L. Goodale, N. A. Jaworski and R. W. Howarth. 2002. Anthropogenic nitrogen sources and relationships to riverine nitrogen export in the northeastern U.S.A. Biogeochemistry 57/58:137-169.

Boyer, E. W., R. W. Howarth, and J. N. Galloway. Riverine nitrogen export from the world’s watersheds. Presentation at the 17th Biennial conference of the Estuarine Research Federation, Seattle, WA, September 14-18, 2003.

Howarth, R.W., R. Marino, E.W. Boyer and D.P. Swaney. Potential consequences of climate change on delivery of nutrients to estuaries. Presentation at the 17th Biennial conference of the Estuarine Research Federation, Seattle, WA, September 14-18, 2003.

Howarth, R.W., A. Sharpley and D. Walker. 2002. Sources of Nutrient Pollution to Coastal Waters in the United States: Implications for Achieving Coastal Water Quality Goals. Estuaries 25:656-676.

Mayer, B., E.W. Boyer, C. Goodale, N. A. Jaworski, N.ico Van Breemen,

R.W. Howarth, S. Seitzinger, G. Billen, K. Lajtha, K. Nadelhoffer, D. Van Dam,

L. J. Hetling, M. Nosal and K. Paustian. 2002. Sources of nitrate in rivers draining sixteen watersheds in the northeastern U.S.: Isotopic constraints. Biogeochemistry 57/58:171-197.

Seitzinger, S. P., R. V. Styles, E. W. Boyer, R. B. Alexander, G. Billen, R. W.

Howarth, B. Mayer and N. Van Breemen. 2002. Nitrogen retention in rivers: model development and application to watersheds in the northeastern U.S.A. Biogeochemistry 57/58:199-237.

Swaney, D.P., R.W. Howarth and E.W. Boyer. Implementing a management oriented nutrient loading model in Excel/VBA. (in preparation for submission to Ecological Modeling)

Swaney, D.P., R. W. Howarth and E.W. Boyer. ReNuMa: A regional scale nutrient loading model for management. Poster presentation at the 17th Biennial conference of the Estuarine Research Federation, Seattle, WA, September 14-18, 2003.

Van Breemen, N., E. W. Boyer, C. Goodale, N. A. Jaworski, K. Paustian, S. P. Seitzinger, K. Lajtha, B. Mayer, D. Van Dam, R. W. Howarth, J.J. Nadelhoffer, M. Eve, and G. Billen. 2002. Where did all the nitrogen go? Fate of nitrogen inputs to large watersheds in the northeastern USA. Biogeochemistry 57/58:267-293.

References

Bricker, S. B., C. G. Clement, D. E. Pirhalla, S. P. Orland, and D. G. G. Farrow. 1999. National Estuarine Eutrophication Assessment: A Summary of Conditions, Historical Trends, and Future Outlook. National Ocean Service, National Oceanic and Atmospheric Administration, Silver Springs, MD.

Dodd, R.C. and J.P. Tippett. 1994. Nutrient Modeling and Management in the Tar- Pamlico River Basin. Research Triangle Institute. Unpublished Report.

Goolsby, D. A., W. A. Battaglin, G. B. Lawrence, R. S. Artz, B. T. Aulenbach, and R. P. Hooper. 1999. Gulf of Mexico hypoxia assessment, Topic #3, Flux and sources of nutrient in the Mississippi-Atchafalaya River basin. Committee on Environmental and Natural Resources, Hypoxia Work Group for the Mississippi River/Gulf of Mexico Watershed Nutrient Task Force

Haith, D.A. and L.L. Shoemaker. 1987. Generalized watershed loading functions for stream flow nutrients. Wat. Res. Bull. 23(3):471-478.

Howarth, R. W. 1988. Nutrient limitation of net primary production in marine ecosystems. Ann. Rev. Ecol. & Syst. 19: 89-110.

Howarth, R.W, J.R. Fruci, and D. Sherman. 1991. Inputs of sediment and carbon to an estuarine ecosystem: Influence of land use. Ecol. Appl. 1:27-39

Howarth, R.W., G. Billen, D. P. Swaney, A. Townsend, N. Jaworski, K. Lajtha, J. A. Downing, R. Elmgren, N. Caraco, T. Jordan, F. Berendse, J. Freney, V. Kudeyarov, P. Murdoch, Zhu Zhao-liang. 1996. Riverine Inputs of Nitrogen to the North Atlantic Ocean: Fluxes and Human Influences. Biogeochemistry, 35:75-139.

Jaworski, N. A., R. W. Howarth, and L. J. Hetling. 1997. Atmospheric deposition of N oxides onto the landscape contributes to coastal eutrophication in the northeast United States. Environ. Sci. Technol. 31: 1995-2004.

Lee, Kuang-Yao, T. R. Fisher, T. E. Jordan, D. L. Correll, and D. E. Weller.

2000. Modeling the hydrochemistry of the Choptank River basin using GWLF

and Arc/Info: 1. Model calibration and validation. Biogeochemistry 49:

143-173.

National Research Council. 1993. Managing Wastewater in Coastal Urban Areas. National Academy Press, Washington, DC.

National Research Council. 2000. Clean Coastal Waters: Understanding and Reducing the Effects of Nutrient Pollution. National Academy Press, Washington, DC.

Nixon, S. W. 1995. Coastal marine eutrophication: a definition, social causes, and future concerns. Ophelia 41: 199-219.

Smith, R. A., G. E. Schwarz, and R. B. Alexander. 1997. Regional interpretation of water-quality monitoring data. Wat. Resour. Res. 33(12):2781-2798.

Swaney, D.P., D. Sherman, and R.W. Howarth. 1996. Modeling Water, Sediment, and Organic Carbon Discharges in the Hudson/Mohawk Basin: Coupling to Terrestrial Sources. Estuaries, 19(4):833-847.

Appendix 1. GWLFXL1.0 source code. October, 2003.

1) Module 1. Model

Option Base 1

' Module-level variable declarations

'*******************************************************************

' DIMENSIONS - MAIN PROG & OUTPUT

'*******************************************************************

Public iter As Integer

Public STREAMFLOW(), PRECIPITATION(), evapotrans(), GROUNDWATER(), RUNOFF() As Single

Public EROSION(), SEDYIELD() As Single

Public GROUNDNITR(), GROUNDPHOS(), DISNITR(), TOTNITR(), DISPHOS(), TOTPHOS() As Single

Public wetdepnitr(), drydepnitr(), lu_denit(), gwdenitfrac, gwdenit(), denitfrac(), denittot()

Public lu_RUNOFF(), lu_EROSION(), lu_DISNITR(), lu_TOTNITR(), lu_DISPHOS(), lu_TOTPHOS() As Single

'*********************************************************************

' DIMENSIONS - MAIN PROG & NUTRIENTS

'*********************************************************************

Public nitrconc(200), phosconc(200), mannitr(200), manphos(200), urbannitr(200), urbanphos(200), pointnitr(0 To 12) As Single, pointphos(0 To 12) As Single

Public x(15, 12), Q(15, 12), SEDTRANS() As Single, BSED(12), URBANSED(200) As Single

Public temp(), PREC(), tempav(), DAYSMONTH(12) As Single

'***********************************************************************

' DIMENSIONS - MAIN PROG, TRANSPORT,NUTRIENTS & 0UTPUT

'***********************************************************************

Public dayhrs(12), grow(12), AREA(200), KLSCP(200), CN(3, 200), ANTMOIST(5) As Single

Public month$(12, 4), landuse$(15, 12), areafinal(200), annareainc(200), AREAINIT(200)

Public dsN(4, 12), dsP(4, 12), a(4, 12), septicN(), septicP()

'declare remaining global variables

Public acoef(12), cv(12), foldername$

Public RECESSCOEF, SEEPCOEF, UNSATSTOR, SATSTOR, snow, SEDELRATIO As Single

Public sednitr, sedphos, grnitrconc, grphosconc As Single

Public rain, melt, water, amc5, qtotal As Single

Public pdry, ndeptot, ndepc() As Single

Public nday() As Integer

Public manuredareas, firstmanuremonth, lastmanuremonth, nrur, nurb, nlu, nyrs As Integer

Public dormantseason As Boolean, growingseason As Boolean, trajectory As Boolean

Public optimize As Boolean, report As Boolean, unsens As Boolean

Public eloadN, eloadP, upN, upP, dsflag, inityr, initmo

Public etmult, erosmult ' multipliers used in calibration

Public gout(120) As Single, outlen As Long, num As Integer 'num = number of params to be optimized

Public nlist(20) As Integer ' row locations of optimization parameters offset by -1

Sub gwlf4(opt, Optional optimize As Boolean, Optional unsens As Boolean)

Dim line1, line2, line3, Title, default, MyValue As String

Dim lenmo(12) As Integer

foldername$ = Worksheets("notes and parameters").Cells(7, 4).value

Do While Not PathExists(foldername$)

foldername$ = getstrg$("Old foldername doesn't exist. New foldername?", foldername$)

If Len(foldername$) = 0 Then

MsgBox " Simulation halted...hit ok to continue", vbInformation

End

End If

Loop

Worksheets("notes and parameters").Activate

If optimize Then

report = False

unsens = False

Else

report = True

Worksheets("notes and parameters").Cells(7, 4).value = foldername$

End If

If unsens Then

report = False

optimize = False

End If

az = 0

If Not optimize Then Worksheets("notes and parameters").Cells(3, 9).value = "Simulation beginning"

default = Str(opt)

numday = 0

monthct = 0

410 If opt > 3 Then End

'************************************************************************

' MAIN PROGRAM

'*************************************************************************

' SIMULATION INITIALIZATION

'*************************************************************************

If Not optimize Then Simbegin.Show

DAYSMONTH(1) = 31

DAYSMONTH(2) = 28

DAYSMONTH(3) = 31

DAYSMONTH(4) = 30

DAYSMONTH(5) = 31

DAYSMONTH(6) = 30

DAYSMONTH(7) = 31

DAYSMONTH(8) = 31

DAYSMONTH(9) = 30

DAYSMONTH(10) = 31

DAYSMONTH(11) = 30

DAYSMONTH(12) = 31

If Not optimize Then

Worksheets("notes and parameters").Cells(1, 9).value = 0

Worksheets("notes and parameters").Cells(2, 9).value = 0

End If

'get parameters

spinuplen = Worksheets("notes and parameters").Cells(5, 4).value

If spinuplen > 0 Then

If UCase$(Worksheets("notes and parameters").Cells(6, 4).value) = "Y" Then

spinuprepeat = True

Else

spinuprepeat = False

End If

End If ' spinuplen > 0

simtitle$ = Worksheets("notes and parameters").Cells(8, 4).value

nyrs = Worksheets("notes and parameters").Cells(9, 4).value

If iter = 1 Or Not unsens Then

ReDim temp(1 - spinuplen To nyrs, 12, 31), PREC(1 - spinuplen To nyrs, 12, 31)

ReDim nday(1 - spinuplen To nyrs), ndepc(1 - spinuplen To nyrs)

Else

ReDim Preserve temp(1 - spinuplen To nyrs, 12, 31), PREC(1 - spinuplen To nyrs, 12, 31)

ReDim Preserve nday(1 - spinuplen To nyrs), ndepc(1 - spinuplen To nyrs)

End If

ReDim STREAMFLOW(1 - spinuplen To nyrs, 0 To 12), PRECIPITATION(1 - spinuplen To nyrs, 0 To 12), evapotrans(1 - spinuplen To nyrs, 0 To 12), GROUNDWATER(1 - spinuplen To nyrs, 0 To 12), RUNOFF(1 - spinuplen To nyrs, 0 To 12) As Single

ReDim EROSION(1 - spinuplen To nyrs, 0 To 12), SEDYIELD(1 - spinuplen To nyrs, 0 To 12) As Single

ReDim GROUNDNITR(1 - spinuplen To nyrs, 0 To 12), GROUNDPHOS(1 - spinuplen To nyrs, 0 To 12), DISNITR(1 - spinuplen To nyrs, 0 To 12), TOTNITR(1 - spinuplen To nyrs, 0 To 12), DISPHOS(1 - spinuplen To nyrs, 0 To 12), TOTPHOS(1 - spinuplen To nyrs, 0 To 12) As Single

ReDim wetdepnitr(1 - spinuplen To nyrs, 0 To 12), drydepnitr(1 - spinuplen To nyrs, 0 To 12)

ReDim lu_RUNOFF(1 - spinuplen To nyrs, 200), lu_EROSION(1 - spinuplen To nyrs, 200), lu_DISNITR(1 - spinuplen To nyrs, 200), lu_TOTNITR(1 - spinuplen To nyrs, 200), lu_DISPHOS(1 - spinuplen To nyrs, 200), lu_TOTPHOS(1 - spinuplen To nyrs, 200) As Single

ReDim SEDTRANS(1 - spinuplen To nyrs, 0 To 12) As Single

ReDim septicN(1 - spinuplen To nyrs, 0 To 12), septicP(1 - spinuplen To nyrs, 0 To 12)

ReDim tempav(1 - spinuplen To nyrs, 0 To 12), denittot(1 - spinuplen To nyrs, 0 To 12)

ReDim gwdenit(1 - spinuplen To nyrs, 0 To 12), lu_denit(1 - spinuplen To nyrs, 0 To 12)

If nyrs = 0 Then Stop

Call initializeall(nyrs)

If UCase$(Worksheets("notes and parameters").Cells(15, 4).value) = "Y" Then trajectory = True Else trajectory = False

If UCase$((Worksheets("notes and parameters").Cells(2, 4).value) = "F") Then

Call readtransportfile(foldername$ & Worksheets("notes and parameters").Cells(27, 4).value) ' transport filename

Call clearworksheet("transport")

Call reporttransporttoworksheet("transport", 0, 0)

Call readnutrientfile(foldername$ & Worksheets("notes and parameters").Cells(28, 4).value) ' nutrient filename

Call clearworksheet("nutrient")

Call reportnutrienttoworksheet("nutrient", "watershed", 1, 0, 0)

Else

Call readtransportworksheet("transport")

Call readnutrientworksheet("nutrient")

End If

zzzz = denitfrac(4)

' use latitude to calculate dayhrs?

latyn$ = Worksheets("notes and parameters").Cells(16, 4)