understand sources of N delivered to US coastal areas Michelle - - PowerPoint PPT Presentation

Using model comparisons to understand sources of N delivered to US coastal areas

Michelle McCrackin Baltic Eye Stockholm University Baltic Sea Centre 18 May 2015 michelle.mccrackin@su.se

• Recommendation: 25%

reduction in N inputs

• Identifying N sources

– Critical to achieve load reductions – Measuring source attribution is challenging – Limitations to isotope analyses

Context for analysis

NASA Earth Observatory

Trends in GoM

NASA Earth Observatory

Alexander et al. 2002

16 NE US watersheds

Past Model Comparisons

1 watershed in Estonia

Lidén et al. 1999

17 watersheds in Europe EuroHarp

Multitude of Models

HSPF: Hydrologic Simulation Program Fortran NANI: Net Anthropogenic Nitrogen Inputs NEWS: Nutrient Export from Watersheds NLM: Nitrogen Loading Model SWAT: Soil Water Assessment Tool SPARROW: Spatially Referenced Regressions On Watersheds WATERSN: Watershed Assessment Tool for Evaluating Reduction Strategies for Nitrogen Others…

5

Multitude of Models

HSPF: Hydrologic Simulation Program Fortran NANI: Net Anthropogenic Nitrogen Inputs NEWS: Nutrient Export from Watersheds NLM: Nitrogen Loading Model SWAT: Soil Water Assessment Tool SPARROW: Spatially Referenced Regressions On Watersheds WATERSN: Watershed Assessment Tool for Evaluating Reduction Strategies for Nitrogen Others…

6

Study objectives

Identify:

• important sources of N to US

coasts

• areas of agreement/disagreement

between NEWS and SPARROW

• factors that contribute to

differences in predictions

Model overview

NEWS SPARROWUS

Description

Combines statistical and mechanistic approaches; simple, can be run in Excel Hybrid statistical and mechanistic approach

Parameters

2 calibrated parameters estimated at global scale 11 calibrated parameters estimated at reach scale Non-significant predictors excluded

Scale

Applied globally, 30-minute resolution; US: 170 basins Applied at US national and regional scale; US: 7,000 US basins

Comparison of predictions

Comparison of yield predictions

McCrackin et al. 2013

NEWS predictions:

• within factor of 2 of SPARROW

for >50% of catchments

• within factor of 4 for 80%

Human activities drive 70-90% of N delivered to US coastal zone

SPARROWUS NEWS

42% 20% 21% 5% 11% 31% 17% 9% 12% 32%

McCrackin et al. 2013

NEWS and SPARROW agree for 67% modeled area

Single-largest N sources

McCrackin et al. 2013

Single-largest N sources

SPARROWUS NEWS

McCrackin et al. 2013

Single-largest N sources

McCrackin et al. 2013

Factors affecting differences in model predictions

NEWS SPARROWUS

Fertilizer and Manure

Gridded global fertilizer and manure data from UN-FAO Estimated N-fixation on agricultural land (1.9 Tg/y) County-based fertilizer purchases Manure estimated from livestock census Estimated N-fixation by key crops (7.3 Tg/y)

Atmospheric deposition

Modeled global TN deposition

(Dentener 2006)

Wet NO3

• deposition (NADP)

Sewage

Human sewage as function of GDP Human population

Natural N

Estimated N-fixation on non- agricultural land (Cleveland et al.

1999)

Surface area of forest, shrub, barren land

Data inputs

Uncertainties in input data

Sobota et al. 2013

National-level estimates

Model structure

NEWS SPARROWUS

Terrestrial N retention

Function of runoff Function of temperature, soil permeability, stream density

Aquatic N retention

Denitrification, dams, irrigation water withdrawals First-order decay

Baltic Sea applications?

Riverine sources to the Baltic

Arheimer et al., 2013 Balt-HYPE PLC 5

45% 16% 20% 19% Diffuse/Agricul ture Background Point/Sewage Other

Riverine N and P to the Baltic

TN TP

Balt-HYPE