201• 529 • 5151www.hydroqual.com

Use of Mechanistic Modeling to Enhance Derivation of Great Bay TN Criteria and Inform

Restoration Strategy

Thomas W. Gallagher, P.E.Cristhian A. Mancilla, EIT

Use of Mechanistic Modeling to Enhance Derivation of Great Bay TN Criteria and Inform

Restoration Strategy

Thomas W. Gallagher, P.E.Cristhian A. Mancilla, EIT

Southeast Watershed Alliance SymposiumMay 11, 2011

Stressor-Response Regression Analysis

Stressor-Response Regression Analysis

Regresses response variable (high chla, low DO, eelgrass decline) to stressor (nutrients). MAJOR ASSUMPTION: Other factors not significantly influencing “dose:response”

Empirical relationship (Vollenweider and others) of lake chla to lake phosphorus have been successfully applied for over 40 years.

Stressor-response relationships for rivers and estuaries have been more problematic.

Primary difficulties have been the proper consideration of other factors that influence the stressor-response relationship and the covariance across some stressors.

Mechanistic ModelingMechanistic Modeling

Relates response variables to stressor with consideration of site specific physical, chemical, and biological factors.

Example: Low DO due to algal photosynthesis and respiration, BOD oxidation, nitrification, sediment oxygen demand and atmospheric reaeration.

Requires a considerable amount of data in comparison to the stressor-response approach.

More expensive than stressor-response approach but yields more scientifically defensible results; however, frequently beyond the budget of many regulatory agencies.

EPA Science Advisory Board Stressor Response Observations

EPA Science Advisory Board Stressor Response Observations

“In order to be scientifically defensible, empirical methods must take into consideration the influence of other variables….  The statistical methods in the Guidance require careful consideration of confounding variables before being used as predictive tools. … Without such information, nutrient criteria developed using bivariate methods may be highly inaccurate.”

SAB Stressor Response Review – April 27, 2010

Proposed Numeric Nitrogen Criteria for Great Bay Estuary

Proposed Numeric Nitrogen Criteria for Great Bay Estuary

Trend Monitoring Stations for Water Quality in the Great Bay Estuary

(New Hampshire DES, 2009)

Trend Monitoring Stations for Water Quality in the Great Bay Estuary

(New Hampshire DES, 2009)

Relationship between Light Attenuation Coefficient and TN at Trend Stations

(New Hampshire DES, 2009)

Relationship between Light Attenuation Coefficient and TN at Trend Stations

(New Hampshire DES, 2009)

0.75

Relationship between Turbidityand TN at Datasonde Stations

(New Hampshire DES, 2009)

Relationship between Turbidityand TN at Datasonde Stations

(New Hampshire DES, 2009)

Contributions to Kd (PAR) measured at the Great Bay Buoy

(From Morrison et al, 2008)

Contributions to Kd (PAR) measured at the Great Bay Buoy

(From Morrison et al, 2008)

Measured Chla and Secchi Disk at Adams Point (1988-2009)

Measured Chla and Secchi Disk at Adams Point (1988-2009)

Relationship between Minimum DO and Chlorophyll-a at Trend Stations

(New Hampshire DES, 2009)

Relationship between Minimum DO and Chlorophyll-a at Trend Stations

(New Hampshire DES, 2009)

Coastal

Bay

Tidal River

Other factors: - Residence time - Reaeration - SOD - BOD oxidation -Stratification

DO at the Squamscott River DatasondeJuly 2008

DO at the Squamscott River DatasondeJuly 2008

DO (%)

DO (mg/L)

Depth

Salinity

DO at the Squamscott River DatasondeJuly-October 2009

(New Hampshire DES, 2011)

DO at the Squamscott River DatasondeJuly-October 2009

(New Hampshire DES, 2011)

Factors InfluencingWater Column Dissolved Oxygen

Factors InfluencingWater Column Dissolved Oxygen

BOD5 DO

Reaeration

Algal photosynthesis and respiration

SOD

NH4

BOD5

Water column stratification

Nitrogen Phosphorus

Residence Time

RecommendationsRecommendations

Continue to look for other factors responsible for eelgrass decline other than changes in water column light transparency.

To better understand the factors responsible for low DO in tributaries to Great Bay, perform a data collection effort to support mechanistic modeling.

Start with Squamscott River DO model to address Exeter permit issues.

STOP

Relationship between TN and Chlorophyll-a at Trend Stations

(New Hampshire DES, 2009)

Relationship between TN and Chlorophyll-a at Trend Stations

(New Hampshire DES, 2009)

COASTAL MARINE LABORATORY

GREAT BAY DATASONDE

LAMPREY RIVER DATASONDE

90th chla = 9.3 ug/LMedian TN = 0.39 mg/L

90th chla = 7.5 ug/LMedian TN = 0.45 mg/L

SALMON FALLS RIVER DATASONDE

OYSTER RIVER DATASONDE

SQUAMSCOTT RIVER DATASONDE

90th chla = 14.3 ug/LMedian TN = 0.57 mg/L

90th chla = 12.1 ug/LMedian TN = 0.74 mg/L

90th chla = 13.7 ug/LMedian TN = 0.52 mg/L

Daily Minimum DO, Jun-Sep 2000-2008 (New Hampshire DES, 2009) Daily Minimum DO, Jun-Sep 2000-2008 (New Hampshire DES, 2009) 90th chla = 3.3 ug/LMedian TN = 0.30 mg/L

Key Transparency Issues to ResolveKey Transparency Issues to Resolve

Did Bay transparency significantly change over time? Degree of a 1-2 ug/l Chlorophyll a change on long term

average transparency? Degree of chlorophyll a reduction achievable with TN

control considering bay hydrodynamics? Does pattern of eelgrass loss fit transparency theory? Are other factors at play in eelgrass decline?

Summary statistics for DO and chlorophyll-a for grab samples collected from 2000-2008 grouped by assessment zone (NH

DES. 2009. Figure 26)

Predicted threshold for DO violations: Chl-a (90th %ile) > 7 ug/L

Summary statistics for DO and chlorophyll-a for grab samples collected from 2000-2008 grouped by assessment zone (NH

DES. 2009. Figure 26)

Predicted threshold for DO violations: Chl-a (90th %ile) > 7 ug/L

y = -0.1214x + 5.7986R2 = 0.4995

y = 0.2213x + 12.241R2 = 0.4268

0

5

10

15

20

0 5 10 15 20

90th %ile Chlorophyll-a (ug/L)

DO

(m

g/L

)

Min

Max

N>20 for all points