Phosphorus Risk Assessment Using the Variable Source Loading Function Model

Size: px
Start display at page:

Download "Phosphorus Risk Assessment Using the Variable Source Loading Function Model"

Transcription

1 Phosphorus Risk Assessment Using the Variable Source Loading Function Model Zachary Easton 1, Elliot Schneiderman 2, M. Todd Walter 1, and Tammo Steenhuis 1 1 Dept. Biological and Environmental Engineering, Cornell University, Ithaca, NY 2 New York City Dept. Environmental Protection, Kingston, NY * zme2@cornell.edu

2 We know that in many areas with steep slopes and shallow soils hydrology is, to a great extent, driven by topology. Variable Source Areas (VSA) Near stream areas, areas with shallow, slowly permeable soil or a large contributing areas produce the majority of runoff

3 Why is P Modeling Difficult in VSA Watersheds Spatially and temporally heterogeneous Watershed properties (soil, landuse, topology, STP, management) can display great variability Thus another level of complexity is added to predicting P loss How can we as scientists, planners, and watershed managers identify areas of the landscape prone to high runoff and thus potentially high P losses?

4 Many Current Water Quality Models Cannot Capture this Complexity, e.g., General Watershed Loading Function (GWLF) Soil Water Assessment Tool (SWAT) Agricultural Nonpoint Source Pollution Model (AGNPS)

5 First Then we must be delineate able to P correctly source areas predict where runoff source areas are Now we can visualize and begin to manage Only critical distributed source areas models can do this

6 Plant tissue Disturbed un-vegetated sites Impervious surfaces Fertilized areas P Sources Manured Areas

7 Modeling P Loss Since different P sources contribute P differently, model processes separately Fertilizer/manure Impervious areas Baseflow Soils Hydrologic input were taken from VSLF and used in the P model

8 Dissolved P (kg ha -1 ) Fertilizer/manure P Fertilizer/manure P declines with successive runoff (R) events following application: D F,t is the dd F, t dr k M D F F, t F, t Apr-03 Aug-03 Dec-03 Apr-04 Aug-04 Dec-04 Apr-05 2 cumulative P loss, k F is the reaction constant, M F,t is the amount of P left at time t

9 P Load (kg ha -1 ) Impervious Surfaces P accumulates on impervious surfaces Modeled with an exponential relationship Apr-03 May-03 Jun-03 Wash off of accumulated P is modeled with a first order relationship

10 Baseflow Represents integrated influence of surface and subsurface factors. An export coefficient approach was used L B B, t B t where L B,t is baseflow P load, μ B is the mean expected P concentration in baseflow, and B t is the baseflow volume

11 P Runoff (mg L -1 ) Soil P P losses from soil can be approximated using soil test P values: D M R S, t S S t where D S,t is the runoff P load μ s is the soil specific coefficient determined from sampled runoff M S is the soil test P R t is the runoff 0.3 y = x r 2 = Morgan's STP (mg kg -1 )

12 SRP Load (kg/d) Discharge (cm/d) Agricultural

13 TDP Load (kg d -1 ) DP Load (kg d -1 ) 1.4 Developed Measured Modeled Measured Modeled 0.0 Nov-03 Dec-03 Jan-04 Feb-04 Mar Apr-03 Jul-03 Oct-03 Jan-04 May-04 Aug-04 Nov-04 Mar-05

14 SRP (mg/l) Dissolved P (ug L -1 ) How Does Temp Affect P Availability? Suburban Watershed Ithaca, NY Forested Watershed (Catskills, NY) Scott et al Biogeochemistry Apr-03 Feb-96Jul-03 Jun-96 Oct-03 Jan-04 Oct-96 May-04 Aug-04 Feb-97 Nov-04 Jun-97

15 Phosphate (mg/l) 0.8 Possible Link to DOC R 2 = DOC (mg/l)

16 DOC in Streamflow (mg/l) Simulated Linking Biogeochemical Systems DOC P Chile Study 12 0 y = 1.00x R 2 = Observed 12 Van t Hoff (1898) C Q 10 T 10 T o Valdivia, Walter, Hedin <in preparation> 12/01/ /19/ /05/ /24/ /09/1999 Observed DOC Simulated DOC

17 Temp (C) How Does Temp Affect P Availability? Q S T, S S T TR 10 Soil Surface Baseflow (50cm) -20 Apr-03 Jul-03 Oct-03 Jan-04 Apr-04 Jul-04 Oct-04 Jan-05 Apr-05

18 SRP Load (kg/d) Discharge (cm/d) Improved Corroboration Hively et al HESS

19 DP Load (kg d -1 ) TDP Load (kg d -1 ) DP Load (kg d-1) Streamflow (mm d -1 ) Measured Modeled Urban Measured 18.3 Modeled 19.5 Measured 23.9 Modeled Apr-03 Jun-03 Sep Dec-03 Mar-04 Jun-04 Sep-04 Dec-04 Mar Measured Modeled r 2 = 0.77 E = 0.75 Measured1 Modeled r 2 = 0.77 E = 0.75 Measured Modeled Nov-03 Dec-03 Jan-04 Feb-04 Mar-04 Measured 3.2 Modeled Apr-03 Jul-03 Oct-03 Jan-04 May-04 Aug-04 Nov-04 Mar-05 Apr-03 Jul-03 Oct-03 Jan-04 May-04 Aug-04 Nov-04 Mar-05

20 Where are the High Risk Areas? High Runoff Low P High Runoff High P Low Runoff High P Low Runoff Low P Weti P Loss LU Dissolved (kg/ha) P 25-Jan-1999 Sept

21 P Loss (kg/ha)

22 VSAs Discussed Spatial and temporal components to P loss Areas with high runoff losses can be P source areas, but not all of them! Interaction between runoff source areas and P source areas

23 Conclusions Landscape P loss originates from definable areas of the watershed VSAs Landscapes contribute differing P loads at different times.temp dependency Winter vs. summer Soils vs. fertilized/manured vs. baseflow To reduce P loads in surface waters we need to consider all these factors