Phosphorus index project overview collick

Phosphorus Index Project Overview: Refining and Harmonizing Phosphorus Indices in the Chesapeake Bay Region to Improve Critical Source

Area Identification and to Address Nutrient Management Priorities

Presented by Amy S. CollickAgriculture, Food, and Resource SciencesUniversity of Maryland Eastern Shore (UMES)

CollaboratorsUniversity of Maryland Eastern Shore

A. Allen, A. CollickPasture Systems and Watershed Management Research Unit, USDA‐ARS

P. Kleinman, T. Veith, R. Bryant, T. Buda, J. Liu, M. AminPennsylvania State University

D. Beegle, J. WeldVirginia Tech

Z. Easton, D. Fuka, M. ReiterCornell University

Q. Ketterings, K. Czymmek, S. Cela, S. CrittendenUniversity of Delaware

A. Shober, K. Clark, S. TingleWest Virginia University

T. BasdenUniversity of Kentucky

J. McGrath

Spring Creek

MahantangoCreek

ConewagoCreek

Allegheny PlateauDressler

Run

Nanticoke River(Bucks Branch)

Upper Manokin

River

Factory Brook

Valley & Ridge

Piedmont

Coastal Plain

Shenandoah River

Chesapeake Conservation Innovation Grant Watersheds

1. Model proving

Monitoring database

3. Site assessment tool

evaluation

Project Approach

Phosphorus routines

Alternative models

Lessons and Achievements

• Model Improvement: Better representation of critical source areas and nutrient transport

• Greater understanding of challenges/limitations in modeling P risk

• Model component comparisons (i.e., erosion) to pinpoint where changes are likely needed in P index

• Informed evaluations and modifications by nutrient management planners and stakeholders

• Expanded and enhanced collaborations between partners







Valley & Ridge

This is what we must represent

177

144 44

1

<1

Soil P – mg kg-1

Runoff – litersP loss – kg P ha-1 yr-1

92

Buda et al. JEQ, 2009

Lowest field is now a CREP buffer that

continues to yield largest P loads

4620

DPDPDP

8

DPDPDP78

Hydrologic Routine Testing – Mahantango Creek

• Similar outlet discharge hydrographs

• Better spatial distribution of runoff with TopoSWAT

• Improved identification of nutrient sources with TopoSWAT

Standard SWAT

WE38 outletTopoSWAT

Collick et al., 2014

Detailed field‐scale modeling

https://i.ytimg.com/vi/wx0SJM7FeEc/mqdefault.jpg

http://www.extension.org/sites/default/files/w/4/4a/Spreading_manure.jpg

• Pre‐process SWAT to field‐scale• Every field can have a different set of

practices• Field specific management scenarios

Application rate

Collick et al. JEQ, 2016

6000 gal ac‐1

9000 gal ac‐1

NewOld NewOld

6

3

0

Runoff total Pkg ha‐1

New

Old

Runoff total Pkg ha‐1

1/15 1/31 2/14

0.1

0

0.26000 gal ac‐1

Field 1 Field 2

P routine updates are strongly suggested for

models

Application timing

0.00

0.25

0.50

0.75

1.00

1/1/2010 2/20/2010 4/11/2010

0.0

1.0

2.0

3.0

4.0

1/1/2010 2/20/2010 4/11/2010

01020304050

MeasuredOldNew

No manure applied

Poultry litter applied, January (2 tons ac‐1)

Runo

ff P, m

g L‐1

Rainfall, mm

Runo

ff P, m

g L‐1

Next generation: Forecasting modelsSWAT with weather forecasterWatershed stakeholder decision

support system architecture

• Forecast runoff risks (6‐hrs to 3 days) across Chesapeake Bay watershed

• Provide information for land management decision‐making to reduce nonpoint source pollution risks

• Enabled for smart phones and other GPS‐enabled devices

Sommerlot et al., 2016 Env. Model. & Soft.

Our Solution

• Standardize data access and base datasets using a Broker

Next generation: Data Brokering

Fuka et al., 2016 Env. Model. & Soft.

Data Brokering in ArcSWAT







Upper Manokin River, MD

Coastal Plain

Flat terrain in ditched crop fields, < 1%

4.5

5.0 5.0

5.0

5.5

4.0 4.5

5.0

5.5

Ditches

Monitoring flumes

Contours, 0.5m

0 40 80 120 160 20020Meters

0 40 80 120 160 20020Meters

0 40 80 120 160 20020Meters

0 40 80 120 160 20020Meters

Figure : The delineations of the four ditches on the 1m LiDAR (A), 3m NED DEM (B) and 10m NED DEM (C). Only three drainage areas are apparent from the 10m DEM.

A. 1m DEM B. 3m DEM

C. 10m DEM

0 1 2 30.5Kilometers

'4

0 1 2 30.5Kilometers

HUC12 Manokin River –Taylor Branch

10m DEM

3m DEM

UMES Experimental Farm

D. Manokin Watershed

Ditch 5 Ditch 7

Ditch 6

Ditch 8

Ditch 5 Ditch 7

Ditch 6

Ditch 8

Ditch 5Ditch 6 & Ditch 7

Ditch 8

19

Water Balance Modeling Efforts• Early effort to apply SWAT to the Manokinwatershed at the channel scale at the UMES farm using these datasets to parameterize the model initialization and corroborate the model results

• However, the processes occurring on this flat terrain, the discrepancy of flow in the field channels could not be adequately represented in SWAT

• Potential to couple water balance model with nutrient cycling and transport model

General field layout with field ditches and a schematic of the water balance model

View from above Lateral view

The model (Collick et al., 2006) seems appropriate for Manokin fields (the low slope gradient may still cause problems), which are divided by field drainage ditches that flow into larger public drainage association (PDA) channels

Observed and modeled flow comparisons

Testing the coastal plain P IndexDrainage intensity and distance to drains

Majority of P loss occurs in subsurface flow Empirical work

Geophysical techniques to map shallow flow paths

ERI now supported by NIFA fundsCIG proposal pending for subsurface processes

Factory Brook, NY

Allegheny Plateau

Factory Brook – Limited watershed monitoring

• TopoSWAT without calibration• Automation of farm nutrient

management plan data• APLE model comparison with a

large farm field dataset from across New York

Management Scenarios: Manure timing and method comparisons

0

1

2

3

4

5

6

7

8

9Ap

r‐08

Jul‐0

8

Oct‐08

Jan‐09

Apr‐09

Jul‐0

9

Oct‐09

Jan‐10

Apr‐10

Jul‐1

0

Oct‐10

Jan‐11

Apr‐11

Jul‐1

1

Oct‐11

Jan‐12

Apr‐12

Jul‐1

2

Oct‐12

Dissolved

P, kg ha

‐1

Spring, 6000, surface apply

Spring, 6000, incorporation within 1 day

Fall, 6000, surface apply

Dissolved P loss

Particulate P loss

% %Comparison of application method, same amount of P applied

Spring, 6000, surface apply 100 100Spring, 6000, incorporation within 1 day 66 95Spring, 6000, injection 31 85Comparison of surface versus incorporation and injection, same N supply

Spring, 15500, surface apply 100 100Spring, 6000, incorporation within 1 day 23 52Spring, 6000, injection 11 47Comparison for cover crop use

Fall, 6000, surface apply 100 100Fall, 6000, surface apply + cover crop 110 81

Relative differences between methods to be compared to P index coefficients

Ketterings et al., 2016 in review, JEQ

Spatial Comparison of P index and TopoSWAT

NY Dissolved Pindex, 2009

NY Particulate Pindex, 2009

Low

Medium

High

Very High

Soluble P in TopoSWAT, 2009

Particulate P in TopoSWAT, 2009







Scenario developmentTesting the P Index and SWAT

• Assess range of conditions in physiographic provinces• Simulate management scenarios on range of representative sites

SWAT FrameworkPennsylvania P index

.

. . .. . ....

. .

. . .. . ....

. .

. . .. . ....

. .

. . ... ... ..

..

...

. . .. . ....

. .

. . .. . ....

. .

. . .. . ....

. .

. . ... ... ..

..

...

. . .. . ....

. .

. . .. . ....

. .

. . .. . ....

. .

. . ... ... ..

..

..

.

. . .. . ....

. .

. . .. . ....

. .

. . .. . ....

. .

. . ... ... ..

..

..

.

. . .. . ....

. .

. . .. . ....

. .

. . .. . ....

. .

. . ... ... ..

..

...

. . .. . ....

. .

. . .. . ....

. .

. . .. . ....

. .

. . ... ... ..

..

...

. . .. . ....

. .

. . .. . ....

. .

. . .. . ....

. .

. . ... ... ..

..

..

.

. . .. . ....

. .

. . .. . ....

. .

. . .. . ....

. .

. . ... ... ..

..

..

P Loss

(Mod

eled

or M

easured) Representative* Scenarios Modeled

Monitored Scenario

. Modeled Scenario

Integrating Modeling and Monitoring

* Major Panel Activity

P Index

NY P index and TopoSWAT – 1st Run

R² = 0.29

R² = 0.57

0

5

10

15

20

25

30

35

40

45

0.0 100.0 200.0 300.0 400.0 500.0 600.0

Soluble P from

Top

oSWAT, kg ha

‐1

Dissolved P index

Corn and alfalfa fields Grass fields

Very preliminary results from uncalibrated TopoSWAT run of Factory Brook

1st Run indicates further vetting of results and rerunning when necessary

Pennsylvania P IndexScreening tool

• Special Protection Watershed• Significant management changes

• STP > 200 ppm Mehlich 3• Distance to Stream < 150 feet

Source factors• Soil Test P (Mehlich 3)• Manure P (Rate, Timing x Method, P Availability Coeff.)• Fertilizer P (Rate, Timing x Method)

Transport factors• Erosion (RUSLE2)• Runoff Potential (Drainage Class)

• Subsurface drainage• Distance to the Stream

Modified connectivity

• Riparian Buffer • Grassed waterway • Direct connection

SSURGO variables

SlopeDrainage classesSoil texture classDistance from

streamKsat

AWCTI classOM

USLE KMUKey

Developing reasonable scenariosWhat important conditions are missing in

our watersheds?Field management

Soils

Field delineation

Landuse

Topography

Watershed

Soil texture at variable distance

from stream

SSURGO data -- clay

% Clay







Allegheny PlateauNew York and Pennsylvania

Ridge and Valley/PiedmontPennsylvania and West Virginia Coastal Plain

Delaware

Assess opinions regarding…– Current P Index factors (importance and reliability)– P Index modifications (boundaries and screening tool)

Evaluation and Revision of Phosphorus IndicesQuestionnaire for Nutrient Management Experts

Ketterings et al., 2016 in press, JSWCS







Thank you! Questions??Amy S. Collick [email protected]

Phosphorus index project overview collick

Environment