Draft Phase 6 Watershed Model Updates Modeling Workgroup Conference Call May 2017

05/18/2017 Draft Phase 6 Watershed Model Updates Modeling Workgroup Conference Call May 2017 Gopal Bhatt 1 and Gary Shenk 2 1 Penn State, 2 USGS

Presentation Outline Brief overview of the Phase 6 draft calibration Refinements in the calibration methods Comparison of simulated vs. WRTDS loads Geographic efficiencies of the simulation Simulated monthly loads RIM loads 2

P6 draft River water quality calibration On the May 4 conference call, the improvements in the the Phase 6 draft hydrology and sediment calibrations were reviewed http://www.chesapeakebay.net/calendar/event/25078/ The revised draft inputs, hydrology, and sediment were used for the river water quality calibration. The calibration methods were revised to improve the representation of nitrate and nitrogen as well as phosphorus transport processes. 3

Key versions of river water quality calibrations The first draft was based on revised inputs and model refinements. DRAFT A To improve the nitrate and total nitrogen balance, a regression model was developed and incorporated to simulate nutrient spiraling processes. To improve the riverine sediment and phosphorus transport, the concentration quantiles used in the model calibration were updated. Some of the model parameters (initial, minimum, and maximum) were updated, as well as limited hand calibration was performed. DRAFT G, Draft P6 4

Key versions of river water quality calibrations The first draft was based on revised inputs and model refinements. To improve the nitrate and total nitrogen balance, a regression model was developed and incorporated to simulate nutrient spiraling processes. To improve the riverine sediment and phosphorus transport, the concentration quantiles used in the model calibration were updated. Some of the model parameters (initial, minimum, and maximum) were updated, as well as limited hand calibration was performed. 5

Fate and transport of nitrate and nitrogen A data point corresponds to one of the 77 WRTDS sites. 6

Fate and transport of nitrate and nitrogen WRTDS Nitrate (Per Acre Load) 64.00 32.00 16.00 8.00 4.00 2.00 1.00 0.50 0.25 y = 3E-05x 4-0.0021x 3 + 0.0534x 2 + 0.4093x - 0.2049 R² = 0.98991 y = 0.8786x - 1.2366 R² = 0.9874 0.13 0.13 0.25 0.50 1.00 2.00 4.00 8.00 16.00 32.00 64.00 WRTDS Total Nitrogen (Per Acre Load) 7

PHASE 5 NITRATE 8

BETA 4 NITRATE 9

DRAFT C Revised inputs, model refinements, and parametrization NITRATE 10

DRAFT G With a regression model for land/small-stream processes NITRATE 11

Key versions of river water quality calibrations The first draft was based on revised inputs and model refinements. To improve the nitrate and total nitrogen balance, a regression model was developed and incorporated to simulate nutrient spiraling processes. To improve the riverine sediment and phosphorus transport, the concentration quantiles used in the model calibration were updated. Some of the model parameters (initial, minimum, and maximum) were updated, as well as limited hand calibration was performed. 12

Transport of sediment and phosphorus 16384 147.46 1.0 8192 4096 2048 73.73 36.86 18.43 0.9 0.8 Daily WRTDS Observations Number of Days 1024 512 256 128 64 32 16 9.22 4.61 2.30 1.15 0.58 0.29 0.14 Percent of Days Cummulative Distribution 0.7 0.6 0.5 0.4 0.3 8 4 2 0.07 0.04 0.02 0.2 0.1 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Percent of total load 0.01 0.0 0.001 0.01 0.1 1 10 Total Phosphours (mg/l) 13

Transport of sediment and phosphorus Scour was limited to the top 2% flow events. The erodibility and bed concentration parameters were calibrated based on the biases in the top 5% of the observation samples. 14

15

1E3 1E4 1E5 1E6 1E7 1E8 1E9 PHASE 5 Phase 5.3.2 NITROGEN 30 ESTIMATOR sites (with regional factors) 1E3 1E4 1E5 1E6 1E7 1E8 1E9 16

1E3 1E4 1E5 1E6 1E7 1E8 1E9 BETA 4C Beta 4 calibration NITROGEN 77 WRTDS sites 1E3 1E4 1E5 1E6 1E7 1E8 1E9 17

1E3 1E4 1E5 1E6 1E7 1E8 1E9 DRAFT A Revised inputs and model refinements NITROGEN 77 WRTDS sites 1E3 1E4 1E5 1E6 1E7 1E8 1E9 18

1E3 1E4 1E5 1E6 1E7 1E8 1E9 DRAFT G Revised inputs, model refinements, and calibration methods NITROGEN 77 WRTDS sites 1E3 1E4 1E5 1E6 1E7 1E8 1E9 19

1E3 1E4 1E5 1E6 1E7 1E8 PHASE 5 Phase 5.3.2 PHOSPHORUS 30 ESTIMATOR sites (with regional factors) 1E3 1E4 1E5 1E6 1E7 1E8 20

1E3 1E4 1E5 1E6 1E7 1E8 BETA 4C Beta 4 calibration PHOSPHORUS 61 WRTDS sites 1E3 1E4 1E5 1E6 1E7 1E8 21

1E3 1E4 1E5 1E6 1E7 1E8 DRAFT A Revised inputs and model refinements PHOSPHORUS 61 WRTDS sites 1E3 1E4 1E5 1E6 1E7 1E8 22

1E3 1E4 1E5 1E6 1E7 1E8 DRAFT G Revised inputs, model refinements, and calibration methods PHOSPHORUS 61 WRTDS sites 1E3 1E4 1E5 1E6 1E7 1E8 23

1E4 1E5 1E6 1E7 1E8 1E9 1E10 1E11 PHASE 5 Phase 5.3.2 SEDIMENT 30 ESTIMATOR basins 1E4 1E5 1E6 1E7 1E8 1E9 1E10 1E11 24

1E4 1E5 1E6 1E7 1E8 1E9 1E10 1E11 BETA 4C Beta 4 calibration SEDIMENT 60 WRTDS basins 1E4 1E5 1E6 1E7 1E8 1E9 1E10 1E11 25

1E4 1E5 1E6 1E7 1E8 1E9 1E10 1E11 DRAFT A Revised inputs and model refinements SEDIMENT 60 WRTDS basins 1E4 1E5 1E6 1E7 1E8 1E9 1E10 1E11 26

1E4 1E5 1E6 1E7 1E8 1E9 1E10 1E11 DRAFT G Revised inputs, model refinements, and calibration methods SEDIMENT 60 WRTDS basins 1E4 1E5 1E6 1E7 1E8 1E9 1E10 1E11 27

Review of geographic efficiencies WRTDS and simulated per acre loads are compared. Nash-Sutcliffe model efficiency was used to quantify the predictive power of the model across the watershed. An efficiency of 1 would indicate a perfect match in loads for all river basins (where WRTDS estimates are available). 28

Simulated Per Acre Load Phase 5 geographic efficiencies Nitrate Per Acre Load, NSE = 0.8284 Nitrogen Per Acre Load, NSE = 0.8704 40 35 30 y = 0.7017x R² = 0.97189 40 35 30 y = 0.9166x R² = 0.84429 25 25 20 20 15 15 10 5 10 5 0 0 5 10 15 20 25 30 35 40 0 0 5 10 15 20 25 30 35 40 Phosphorus Per Acre Load, NSE = 0.6321 Sediment Per Acre Load, NSE = -0.077 2.4 2.1 1.8 y = 1.1136x R² = 0.73976 1.6 1.4 1.2 y = 0.0461x R² = -1.3549 1.5 1.0 1.2 0.8 0.9 0.6 0.6 0.4 0.3 0.2 0.0 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 WRTDS Per Acre Load 29

Simulated Per Acre Load Beta 4 geographic efficiencies Nitrate Per Acre Load, NSE = 0.8862 Nitrogen Per Acre Load, NSE = 0.8583 40 35 30 y = 0.8171x R² = 0.91976 40 35 30 y = 1.0165x R² = 0.89273 25 25 20 20 15 15 10 5 10 5 0 0 5 10 15 20 25 30 35 40 0 0 5 10 15 20 25 30 35 40 Phosphorus Per Acre Load, NSE = 0.4497 Sediment Per Acre Load, NSE = 0.3849 2.4 2.1 1.8 y = 0.6694x R² = 0.62279 1.6 1.4 1.2 y = 0.6224x R² = 0.38892 1.5 1.0 1.2 0.8 0.9 0.6 0.6 0.4 0.3 0.2 0.0 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 WRTDS Per Acre Load 30

Simulated Per Acre Load DRAFT A geographic efficiencies Nitrate Per Acre Load, NSE = 0.7918 Nitrogen Per Acre Load, NSE = 0.9225 40 35 30 y = 0.6955x R² = 0.9064 40 35 30 y = 0.9408x R² = 0.93089 25 25 20 20 15 15 10 5 10 5 0 0 5 10 15 20 25 30 35 40 0 0 5 10 15 20 25 30 35 40 Phosphorus Per Acre Load, NSE = 0.5784 Sediment Per Acre Load, NSE = 0.4358 2.4 2.1 1.8 y = 0.7481x R² = 0.64522 1.6 1.4 1.2 y = 0.8948x R² = 0.5018 1.5 1.0 1.2 0.8 0.9 0.6 0.6 0.4 0.3 0.2 0.0 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 WRTDS Per Acre Load 31

Simulated Per Acre Load DRAFT G geographic efficiencies Nitrate Per Acre Load, NSE = 0.9334 Nitrogen Per Acre Load, NSE = 0.9478 40 35 30 y = 0.9909x R² = 0.9347 40 35 30 y = 0.9544x R² = 0.94953 25 25 20 20 15 15 10 5 10 5 0 0 5 10 15 20 25 30 35 40 0 0 5 10 15 20 25 30 35 40 Phosphorus Per Acre Load, NSE = 0.8550 Sediment Per Acre Load, NSE = 0.7423 2.4 2.1 1.8 y = 1.0028x R² = 0.8568 1.6 1.4 1.2 y = 0.9527x R² = 0.71472 1.5 1.0 1.2 0.8 0.9 0.6 0.6 0.4 0.3 0.2 0.0 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 WRTDS Per Acre Load 32

Summary of geographic efficiencies Constituents Phase 5 Beta 4 Draft A Draft G Nitrate 0.8284 0.8862 0.7918 0.9334 Nitrogen 0.8704 0.8583 0.9225 0.9478 Phosphorus 0.6321 0.4497 0.5784 0.8550 Sediment -0.0770 0.3849 0.4358 0.7423 33

Monthly loads: total nitrogen 34

Monthly loads: total phosphorus 35

Monthly loads: sediment 36

RIM loads: total nitrogen BASIN Beta 5G Beta 5A P6 Beta 4 P6 Beta 3 P6 Beta 2 P6 Beta 1 Phase 5 SUSQ -3% -09% -03% -13% 03% -05% -1% POTO -9% -33% -29% -36% -12% -18% -22% JAME -2% -21% -19% -34% -06% -33% -7% RAPP -7% -10% -08% -39% -22% -25% -10% APPO -7% -08% 00% -28% 04% -44% -2% PAMU 0% 00% 05% -19% -04% 07% 6% MATT 8% 24% 37% 16% 13% 57% -6% PATU 6% 10% 21% 00% 53% -03% 19% CHOP 8% 25% 40% 19% 77% 17% -13% assuming +/- 10% uncertainty in WRTDS estimates 37

RIM loads: total phosphorus BASIN Beta 5G Beta 5A P6 Beta 4 P6 Beta 3 P6 Beta 2 P6 Beta 1 Phase 5 SUSQ 14% -14% -11% 12% 02% -08% 10% POTO 0% -45% -49% -14% 04% -36% -14% JAME 10% -40% -46% -23% -27% -54% -25% RAPP 6% -34% -40% -22% 17% -42% 2% APPO 13% -2% -21% 04% 45% -41% 4% PAMU 12% -4% -13% 16% 54% -08% 7% MATT -13% -17% -26% 13% 64% 27% 4% PATU 11% -23% -23% -07% 12% -25% -18% CHOP 13% 5% -10% 32% 151% 147% 22% assuming +/- 15% uncertainty in WRTDS estimates 38

Questions to be asked today Were we successful in our attempt to avoid using calibrated regional factors? Is the calibration without calibrated regional factors as good as the Phase 5 calibration with calibrated regional factors? Would any gain in load accuracy with the addition of calibrated regional factors be worth the loss in explanatory power? 39

Appendices 40

1E3 1E4 1E5 1E6 1E7 1E8 1E9 PHASE 5 Phase 5.3.2 NITRATE 16 ESTIMATOR basins 1E3 1E4 1E5 1E6 1E7 1E8 1E9 41

1E3 1E4 1E5 1E6 1E7 1E8 1E9 BETA 4C Beta 4 calibration NITRATE 77 WRTDS Stations 1E3 1E4 1E5 1E6 1E7 1E8 1E9 42

1E3 1E4 1E5 1E6 1E7 1E8 1E9 DRAFT A Revised inputs and model refinements NITRATE 77 WRTDS Stations 1E3 1E4 1E5 1E6 1E7 1E8 1E9 43

1E3 1E4 1E5 1E6 1E7 1E8 1E9 DRAFT G Revised inputs, model refinements, and calibration methods NITRATE 77 WRTDS Stations 1E3 1E4 1E5 1E6 1E7 1E8 1E9 44

1E2 1E3 1E4 1E5 1E6 1E7 1E8 PHASE 5 Phase 5.3.2 DISSOLVED PHOSPHATE 61 WRTDS Stations 1E2 1E3 1E4 1E5 1E6 1E7 1E8 45

1E2 1E3 1E4 1E5 1E6 1E7 1E8 BETA 4C Beta 4 calibration DISSOLVED PHOSPHATE 61 WRTDS Stations 1E2 1E3 1E4 1E5 1E6 1E7 1E8 46

1E2 1E3 1E4 1E5 1E6 1E7 1E8 DRAFT A Revised inputs and model refinements DISSOLVED PHOSPHATE 61 WRTDS Stations 1E2 1E3 1E4 1E5 1E6 1E7 1E8 47

1E2 1E3 1E4 1E5 1E6 1E7 1E8 DRAFT G Revised inputs, model refinements, and calibration methods DISSOLVED PHOSPHATE 61 WRTDS Stations 1E2 1E3 1E4 1E5 1E6 1E7 1E8 48

Monthly loads: nitrate 49

Monthly loads: dissolved phosphate 50

Annual Phosphorus Runoff Major/Minor Basin Nobs Eastern Shore of Chesapeake Bay 19 Upper Eastern Shore 8 Middle Eastern Shore, including Choptank River 6 Lower Eastern Shore 5 Western Shore of Chesapeake Bay 71 Lower Western shore 71 Patuxent River Basin 9 Patuxent River below Bowie, Maryland 9 Potomac River Basin 10 Lower Potomac River, below Chain Bridge 10 51

SERC data include estimates for phosphorus loads from 36 catchment These catchment intersect with 7 P6 land segments N10005,DE,SUSSEX N24003,MD,ANNE ARUNDEL N24015,MD,CECIL N24017,MD,CHARLES N24033,MD,PRINCE GEORGES N24035,MD,QUEEN ANNES N24045,MD,WICOMICO 52

SERC_1 phosphorus per watershed acres SERC_2 phosphorus per non-natural acres 53