Evaluating the Effectiveness of Stream Restoration in Maryland: Integrating

Existing and New Data from Restoration Monitoring

Chesapeake Bay Trust Forum, June 2019

Solange Filoso

Research Questions

1. How does effectiveness of stream restoration at reducing nutrient and sediment loads vary among different projects implemented in MD?

2. What factors influence project performance?

Potential factors influencing restoration performance

• Stream position in watershed

• Watershed imperviousness

• Catchment size (area)

• Restoration design

• Magnitude of discharge

• Loading concentrations, dominant species of N and P

• Catchment topography, channel slope

Objectives

• To use pre- and post-restoration data from a number of streams in MD monitored with compatible methodology to determine changes in nutrient and sediment loads.

• Calculate effectiveness in terms of removal rates (kg/ha/yr) and relative removal (% of load in pre-restoration stream/site).

• Examine potential factors influencing project performance.

Study StreamsStream

Drainage

area (ha)

Impervious

ness (%)

Position in

watershedWatershed

Physiographic

region

New Data

Collected

Dividing Cr. 89 32 Lowland Magothy Coastal Plain YES

Cabin Br.

(Saltworks)49 55 Lowland Severn Coastal Plain YES

Church Cr. 227 56 Lowland South Coastal Plain YES

Cypress Cr. 143 46 Lowland Magothy Coastal Plain YES

Howard’s Br. 96 11 Lowland Severn Coastal Plain NO

Wilelinor 106 48 Lowland South Coastal Plain NO

Linnean 13 27 Headwater Rock Cr. Piedmont NO

Park Drive 1.3 18 Headwater Anacostia Coastal Plain NO

Clements Cr.

(C. Hills)6 15 Headwater Severn Coastal Plain NO

Red Hill Br. 18 - Headwater Patuxent Coastal Plain NO

Streams monitored

for this project

Streams with

monitoring data

available

Monitoring Designs

Control = Before

Before-After restoration

Control = Above

Above-Below restoration

Paired-catchments

Control = Similar catchment

Before/After/Control/Impact (BACI)

Control = Similar catchment before and after

Above-Below, Before-After restoration

Methods used for data collection: Rain, discharge and water quality

Measuring rain and discharge

Water sampling

Important monitoring attributes

• Data were collected for at least 1 year before and 1 year after restoration.

• Data were collected during base flow and stormflow conditions.

• Base flow samples were collected monthly to quarterly.

• Stormflow samples were collected for at least 8 storm events per year in each site.

• Rain depths and stream flow were recorded continuously.

• Rain data were collected on site for each catchment.

Restored streams were effective at reducing most pollutant loads but reductions varied

LOAD

INCREASED

LOAD

DECREASED

-100

-50

0

50

100

% C

hange

% Change in pollutant loads with restoration

TSS TN TP

Effectiveness of headwater channels was higher than of lowland channels

76 to

91%

0 to

49%

36 to

82%

0 to

34%

TSS TN TP

51 to

76%

0%

Effectiveness was correlated with catchment size

% L

oa

d r

ed

uctio

n

Catchment size (hectares)

Effectiveness decreased with increased imperviousness in catchment

% L

oad

re

du

ctio

n

% imperviousness

Stormflow was a dominant component of annual discharge in smaller and more impervious catchments

0

50

100

PARK DR. C HILLS LINNEAN RED HILL WIL SALT HB DIVIDING CYPRESS CHURCH

%

Base Flow Stormflow

0

50

100

PARK DR. C HILLS LINNEAN RED HILL WIL SALT HB DIVIDING CYPRESS CHURCH

%

Base flow Stormflow

CONTROL/PRE-

RESTORED/POST-

Stormflow concentrations decreased substantially in headwater channels

HEADWATERS LOWLANDS

CTRL

REST

HEADWATERS LOWLANDS

HEADWATERS LOWLANDS

CTRL

REST

HEADWATERS LOWLANDS

CTRL

REST

HEADWATERS LOWLANDS HEADWATERS LOWLANDS

Concentr

atio

n (

mg/L

)F

WM

Concentr

atio

n (

mg/L

)

Restored stream capacity to reduce loads in stormflow tend to decrease with rain size

Example from a lowland channelExample from a headwater channel

Rain events > 1 in were rare but contributed to ~ half of the total annual rain in catchments

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8/7

/11

9/2

/11

9/2

8/1

11

0/2

4/1

11

1/1

9/1

11

2/1

5/1

11

/10

/12

2/5

/12

3/2

/12

3/2

8/1

24

/23

/12

5/1

9/1

26

/14

/12

7/1

0/1

28

/5/1

28

/31

/12

9/2

6/1

21

0/2

2/1

21

1/1

7/1

21

2/1

3/1

21

/8/1

32

/3/1

33

/1/1

33

/27

/13

4/2

2/1

35

/18

/13

6/1

3/1

37

/9/1

38

/4/1

38

/30

/13

9/2

5/1

31

0/2

1/1

31

1/1

6/1

31

2/1

2/1

31

/7/1

42

/2/1

42

/28

/14

3/2

6/1

44

/21

/14

5/1

7/1

46

/12

/14

7/8

/14

8/3

/14

8/2

9/1

49

/24

/14

10/2

0/1

41

1/1

5/1

41

2/1

1/1

41

2/2

8/2

01

41

/23

/201

52

/18

/201

53

/16

/201

54

/11

/201

55

/7/2

015

11/7

/201

61

2/3

/201

61

2/2

9/2

01

61

/24

/201

72

/19

/201

73

/17

/201

74

/12

/201

75

/8/2

017

6/3

/20

17

6/2

9/2

01

77

/25

/201

78

/20

/201

79

/15

/201

71

0/1

1/2

01

71

1/6

/201

71

2/2

/201

71

2/2

8/2

01

7

Daily rain

2011 2012 2013 2014 2015 2016 2017

RAIN SIZE FREQUENCY CONTRIBUTION TO TOTAL VOLUME

Rain

depth

(in

)

Summary

1. 80-90% of the restored streams examined reduced TSS and TN loads,

but only 40% reduced TP loads.

2. Load reduction was relatively higher in headwater channels.

3. Project performance was associated with stream position in watershed,

% imperviousness, and size of catchment.

4. Stormflow contributed most of the annual discharge in headwater

streams; in lowland channels base flow was important as well.

5. Performance of headwater channels was based on their capacity to

reduce loads in stormflow.

6. Performance of lowland channels was based on their capacity to reduce

loads in both base flow and stormflow.

Final Remarks

1. Despite inferior performance of lowland channels, they can

potentially reduce large loads given their size.

2. Trade-offs associated with lowland channels should be

carefully considered.

3. Other factors are likely to influence restoration performance.

4. Synthesis and evaluation of monitoring data is essential to

improve our capacity to predict the outcomes of restoration

projects as well as to develop more cost-effective monitoring

strategies.

Acknowledgments

FilosoTranslation Slides

What does this mean for me? • The efficacy of stream restoration at reducing nutrient and

sediment loads varies among projects.

• Upland streams restored with RCS are generally more effective than valley channels.

• TN, TP, and TSS were consistently reduced in upland systems, while only TN and TSS were reduced in valley systems, but at lower rates.

• Efficacy is associated with the capacity of streams to retain nutrients and sediments during a wide range of storm sizes.

• Restoration improves retention in upland projects during small and large storms, but ONLY in smaller storms in valley channel projects.

• The frequency of storms > 1 inch was only 9% during the monitoring period but they contributed almost 50% of the total rain volume.

What does this mean for me? What do I take from this if I am a practitioner? • Implement projects in headwater areas where feasible to

maximize nutrient/sediment reductions.

What do I take from this if I am a regulator?• Consider site location as an important factor when

reviewing potential projects.

• Upland stormwater best management practices and upland stream restorations may decrease the effect of high flows on downstream areas (e.g. 2018).