HomeMy Public PortalAboutGI Pilot Final Report - Appendix W-ZAPPENDIX W PILOT MONITORING SEMI-ANNUAL REVIEWS
Review of 2013 Green Infrastructure Monitoring Data for MSD GI Pilot ProgramJanuary 23, 2014
2013 Data Review –GI pilotsData Overview by Site•Amended Soil Sites–Labadie–Warne•Bioretention Sites–Vandeventer–Geraldine•Monroe Rain Garden Site
2013 Data Review –GI pilotsAmended Soil Sites
2013 Data Review: Labadie and Warne Amended Soil Sites (Labadie)
2013 Data Review: Labadie and Warne Amended Soil Sites (Labadie)
2013 Data Review: Labadie and Warne Amended Soil Sites (Warne)
Observation:•The main issue is that the screened interval of piezometers is much deeper than shallow soil amendment depth. 2013 Data Review: Labadie and Warne Amended Soil Sites
2013 Data Review: Labadie and Warne Amended Soil SitesRecommendation:•Pull level sensors, cease monitoring –given the depth difference between the surface amendment and the piezometers, it’s unlikely these instruments will yield useful data.•It is also likely that the shallow soil amendments are not significantly affecting infiltration.
2013 Data Review –GI pilotsBioretention Sites
2013 Data Review: VandeventerBioretention00.511.522.533.544.55‐20‐15‐10‐505106/15/2013 0:00 7/5/2013 0:00 7/25/2013 0:00 8/14/2013 0:00 9/3/2013 0:00 9/23/2013 0:00 10/13/2013 0:00 11/2/2013 0:00 11/22/2013 0:00 12/12/2013 0:00Runoff (cfs)15‐minute Rainfall, Clean Out Level, Flume Level (inches)Flume Level (in)Clean Out Level (in)Rainfall (in)Runoff (cfs)
2013 Data Review: VandeventerBioretention – rainfall events 10/31 – 11/200.10.20.30.40.50.60.70.80.9100.0020.0040.0060.0080.010.0120.0140.01610/30/2013 0:00 10/30/2013 12:00 10/31/2013 0:00 10/31/2013 12:00 11/1/2013 0:00 11/1/2013 12:00 11/2/2013 0:00Runoff (cfs)Flume Flow Norm. (cfs)Flume Flow Norm. (cfs)Runoff (cfs)
2013 Data Review: VandeventerBioretentionObservations:•For 10/31 – 11/2 event, it looked like there was an outflow response, but volumes don’t match.•So this would mean that 4,714 c.f. was infiltrated, but testing shows very low infiltration rate.•Also, we don’t see an appropriate response in the underdrain.ComponentVolumeTotal rainfall9,800 c.f.Total runoff (model) 4,729 c.f.Total outflow15 c.f.
2013 Data Review: VandeventerBioretention – November dataInsert plot for11/19 – 12/2•In the November data, we see an apparent response in the outflow to two rainfall/runoff events. •The shape seems right for outflow from the underdrain, but we don’t see an appropriate response in the underdrain.
2013 Data Review: VandeventerBioretentionObservations:•For 11/21 – 11/23 event, it looked like there was an outflow response.•So this would mean that 1,270 c.f. was infiltrated.•Possible, but again, we don’t see an appropriate response in the underdrain.ComponentVolumeTotal rainfall2,754 c.f.Total runoff (model) 1,484 c.f.Total outflow195 c.f.
2013 Data Review: Geraldine BioretentionInsert plot for full time seriesPeak in outflow without commensurate rainfall.Negative values
2013 Data Review: Geraldine BioretentionInsert plot for11/8 – 12/2
2013 Data Review: VandeventerBioretentionObservations:•Underdrain level sensors does not seem to be responding the way we’d expect.•No way of checking ponding in basins.•No way to verify runoff predictions or outflow data.•Data collected to date do not support calculation of performance metrics as planned in GI Monitoring Protocol.
2013 Data Review: VandeventerBioretentionRecommendations:•Add level sensors to measure level in basins, not just underdrains (could use level sensors from Warne & Labadie).•Test level sensors in underdrain by filling at cleanout (garden hose?).•Test basins overall by filling with fire hose.•Start periodic visual observation during rainfall events.
2013 Data Review –GI pilotsMonroe Rain Garden Site
2013 Data Review: Monroe Rain Garden00.010.020.030.040.050.060.07‐505101520253035406/15/2013 0:00 7/5/2013 0:00 7/25/2013 0:00 8/14/2013 0:00 9/3/2013 0:00 9/23/2013 0:00 10/13/2013 0:00 11/2/2013 0:00 11/22/2013 0:00 12/12/2013 0:00Runoff (cfs)15‐minute Rainfall and Flume Level (inches)Flume Level (in)Rainfall (in)Runoff (cfs)
2013 Data Review: Monroe Rain Garden00.010.020.030.040.050.060.07‐1‐0.500.511.5210/30/2013 0:00 10/30/2013 12:00 10/31/2013 0:00 10/31/2013 12:00 11/1/2013 0:00 11/1/2013 12:00 11/2/2013 0:00 11/2/2013 12:00 11/3/2013 0:00 11/3/2013 12:00 11/4/2013 0:00Runoff (cfs)15‐minute Rainfall, Clean Out Level, Flume Level (inches)Flume Level (in)Clean Out Level (in)Rainfall (in)Runoff (cfs)There seems to be a response in the flume to the rainfall……but the level does not return to zero until 11/5.
2013 Data Review: Monroe Rain GardenObservations:•There appears to be some kind of post‐event drift or interference. Could be debris?Recommendations:•Add measuring tape to flume to allow visual measurement of water depth.•Test rain gardens by filling with hose and collecting measurements, observations.•Conduct site visits during rain events to assess flow condition, clogging, etc.
2013 Data Review –GI pilotsDiscussion, next steps...
2014 Mid-Year Review of
Monitoring Data
for MSD’s LTCP Green
Infrastructure Pilot Program
August 21, 2014
2014 Mid-Year GI Data Review
•Data review by site
–Bioretention sites (Vandeventer, Geraldine)
–Monroe rain garden site
–Amended soil sites (Labadie, Warne)
–Humphrey porous alley
•Brief data summary, discuss issues,
observations, conclusions, recommendations
•Summary findings, observations,
recommendations
2014 Mid-Year GI Data Review
Bioretention Sites
2014 Mid-Year GI Data Review: Bioretention
Sites
Data Summary:
•Continuous radar rainfall at both sites
•Underdrain level and outflow monitored
continuously: 1/1/14 through 6/30/14
•Basin level sensors added, data collection
from 4/1/14 through 6/30/14
2014 Mid-Year GI Data Review: Bioretention
Sites
Data Summary:
•Data show good capture
•But, there are some issues with the data
Bioretention Site # Rainfall/ Runoff
Events
(Jan. 1 – Jun. 30)
Total
Runoff
(ft3)
Total
Outflow
(ft3)
Overall
Capture
Vandeventer 65 103,240 585 99%
Geraldine 70 22,136 1,463 93%
2014 Mid-Year GI Data Review: Bioretention
Sites
Data Issues:
•Outflow data requires manual adjustment to
avoid losing information due to sub-zero drift
•Data “noise” during some periods
•Outflow response without rain
•Percent capture is extremely high (is it
believable?)
2014 Mid-Year GI Data Review: Bioretention
Sites
Data issue: data drift
2014 Mid-Year GI Data Review: Bioretention
Sites (Vandeventer)
0
1
2
3
4
5
6-1
-0.5
0
0.5
1
1.5
2
2.5
3
3.5
4
11/1 11/16 12/1 12/16 12/31 1/15 1/30 2/14 3/1 3/16 3/31 4/15 4/30 5/15 5/30 6/14 6/29
Outflow - INCHES Runoff - CFS
2014 Mid-Year GI Data Review: Bioretention
Sites (Vandeventer)
0
1
2
3
4
5
6-1
-0.5
0
0.5
1
1.5
2
2.5
3
3.5
4
11/1 11/16 12/1 12/16 12/31 1/15 1/30 2/14 3/1 3/16 3/31 4/15 4/30 5/15 5/30 6/14 6/29
Outflow - INCHES Adjusted Outflow Runoff - CFS
2014 Mid-Year GI Data Review: Bioretention
Sites (Vandeventer)
0
1
2
3
4
5
6
7
8
9
10-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
5/20 5/23 5/26 5/29 6/1 6/4 6/7 6/10 6/13 6/16 6/19 6/22 6/25 Predicted Runoff (cfs)Depth of Outflow in Flume (inches)RAW Outflow Data Adjusted Outflow Data HEC-HMS Predicted Runoff
2014 Mid-Year GI Data Review: Bioretention
Sites
Data issue: data “noise”
2014 Mid-Year GI Data Review: Bioretention
Sites
Data issues - noise:
•Need to “shift” Geraldine data as with
Vandeventer, but there are also some data
periods that don’t make sense - erratic level
readings
•Some judgment needed in “correcting these”,
must be done manually
2014 Mid-Year GI Data Review: Bioretention
Sites (Geraldine)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7-1
0
1
2
3
4
5
6
11/13 11/28 12/13 12/28 1/12 1/27 2/11 2/26 3/13 3/28 4/12 4/27 5/12 5/27 6/11 6/26
OUTFLOW - Inches Adjusted Outflow - inches Runoff - CFS
Data also requires “shifting”,
but there are several periods
of unreliable data.
2014 Mid-Year GI Data Review: Bioretention
Sites
Data issues – outflow without rain:
•Occurred at Geraldine on 6/27
•Level sensor in basin and outflow flume
showed an “event”
•Rain occurred about 20 hours prior, then again
4-6 hours later
2014 Mid-Year GI Data Review: Bioretention
Sites
Data issue: outflow w/out rain, Geraldine, 6/27
Rain before and
after outflow
Outflow
2014 Mid-Year GI Data Review: Bioretention
Sites
Data issue: extremely high capture
2014 Mid-Year GI Data Review: Bioretention
Sites
Data Summary:
•Data show good capture
•Too good to be true?
•Examined individual events…
Bioretention Site # Rainfall/ Runoff
Events
(Jan. 1 – Jun. 30)
Total
Runoff
(ft3)
Total
Outflow
(ft3)
Overall
Capture
Vandeventer 65 103,240 585 99%
Geraldine 70 22,136 1,463 93%
2014 Mid-Year GI Data Review: Bioretention
Sites (Vandeventer)
Summary (January – June 2014):
•24 rainfall events resulting in runoff > 0.1 in
•Model-predicted runoff volumes = 1,118 c.f. to
9,936 c.f.
•Is the project controlling runoff?
–May 31 (largest) runoff event (from model) = 9,936 c.f.
–Outflow volume (from data) = 115 c.f (99% capture)
–What happened to the other 9,821 c.f.?
2014 Mid-Year GI Data Review: Bioretention
Sites (Vandeventer)
-15
-10
-5
0
5
10
15
20
-1
0
1
2
3
4
5
5/31/2014 6:005/31/2014 9:005/31/2014 12:005/31/2014 15:005/31/2014 18:005/31/2014 21:006/1/2014 0:00Rainfall/Level (inches)Runoff/Discharge (cfs)Runoff (cfs)Discharge (cfs)Precip (in)
Clean Out Level (in)Basin Level (in)Clean Out Spillway Threshold (in)
Basin Spillway Threshold (in)
Event ID: 2014-51
Start Date/Time: 5/31/2014 13:00
Total Duration (hrs): 3
Rainfall Depth (in): 1.174
Total Rainfall Volume (ft3): 11080.2
Peak Rainfall Rate (in/hr): 1.852
Time since last event (days): 57.51
Runoff Duration (hrs): 3.5
Runoff Depth (in): 1.053
Total Runoff Volume (ft3): 9936.0
Peak Runoff Rate (in/hr): 1.168
Total Discharge Volume (ft3): 115.1
Peak Discharge Rate (in/hr): 0.023
Event Discharge Volume (ft3): 90.9
AfterEvent Dischrg Vol. (ft3): 24.2
Evt+AftEvt Dischrg Vol. (ft3): 115.1
Total Volume Stored (ft3): 70.1
Abs. Volume Reduction (gal): 73465.1
Relative Total V Reduction (%): 98.8
Abs. Volume Reduction [Event
Discharge] (gal): 73645.9
Relative Total V Reduction [Event
Discharge] (%): 99.1
2014 Mid-Year GI Data Review: Bioretention
Sites (Vandeventer)
May 31 Event:
•Data show cleanout elevation rises to overflow
(slightly and briefly), consistent with low outflow
•But no rise is storage level in basin
–Did all water infiltrate as soon as it entered basin?
–Rate = 9,821 c.f. in 3 hours or about 5’ distributed over
floor of basin
–Infiltration rate would have to be about 50 cm/hr
which is higher than measured at Vandeventer, but on
par with measurements at Geraldine
2014 Mid-Year GI Data Review: Bioretention
Sites
Data issue - high percent capture:
•Could be real, but it would mean high
infiltration rates
•Cleanout sensors show very slow drainage,
which argues against the high infiltration rate.
–Drainage outside of underdrain could be occurring
through unknown preferential pathways
–Filter material around underdrain could be
clogged
2014 Mid-Year GI Data Review: Bioretention
Sites
Data issue - high percent capture:
•Underdrain data show slow drainage (Vandeventer, 5/31)
~4 days
2014 Mid-Year GI Data Review: Bioretention
Sites (Geraldine)
Summary (January – June 2014):
•32 rainfall events resulting in runoff > 0.1 in
•Model-predicted runoff volumes = 123 c.f. to
2,198 c.f.
•Is the project controlling runoff?
–April 3/4 (largest) runoff event (from model) =
2,198 c.f.
–Outflow volume (from data) = 19 c.f
–What happened to the other 2,179 c.f.?
-20
-15
-10
-5
0
5
10
15
20
25
0
0.1
0.2
0.3
0.4
0.5
0.6
4/2/2014 18:004/2/2014 21:004/3/2014 0:004/3/2014 3:004/3/2014 6:004/3/2014 9:004/3/2014 12:004/3/2014 15:004/3/2014 18:004/3/2014 21:004/4/2014 0:004/4/2014 3:004/4/2014 6:004/4/2014 9:00Rainfall/Level (inches)Runoff/Discharge (cfs)Runoff (cfs)Discharge (cfs)Precip (in)
Clean Out Level (in)Basin Level (in)Clean Out Spillway Threshold (in)
Basin Spillway Threshold (in)
Event ID: 2014-32
Start Date/Time: 4/3/2014 02:45
Total Duration (hrs): 21.75
Rainfall Depth (in): 1.911
Total Rainfall Volume (ft3): 2358.6
Peak Rainfall Rate (in/hr): 1.920
Time since last event (days): 0.44
Runoff Duration (hrs): 22.25
Runoff Depth (in): 1.781
Total Runoff Volume (ft3): 2198.3
Peak Runoff Rate (in/hr): 1.425
Total Discharge Volume (ft3): 18.8
Peak Discharge Rate (in/hr): 0.001
Event Discharge Volume (ft3): 17.6
AfterEvent Dischrg Vol. (ft3): 1.1
Evt+AftEvt Dischrg Vol. (ft3): 18.8
Total Volume Stored (ft3): 237.9
Abs. Volume Reduction (gal): 16304.0
Relative Total V Reduction (%): 99.1
Abs. Volume Reduction [Event
Discharge] (gal): 16312.3
Relative Total V Reduction [Event
Discharge] (%): 99.2
2014 Mid-Year GI Data Review: Bioretention
Sites (Geraldine)
2014 Mid-Year GI Data Review: Bioretention
Sites (Geraldine)
April 3/4 Event:
•Data show neither cleanout nor basin stages
reach overflow
•But no rise is storage level in basin
–Infiltration rate would only have to be about 6
cm/hr which is plausible
2014 Mid-Year GI Data Review: Bioretention
Sites
Data issue - high percent capture:
•Underdrain data show slow drainage (Geraldine, 4/3)
~6 days
2014 Mid-Year GI Data Review: Bioretention
Sites
Summary of data issue (extremely high capture):
•High capture rates could be possible, but
infiltration rates would have to be high (test at
Vandeventer show the opposite)
•Drainage from underdrains is slow, possibly
indicating low infiltration rates
–Could indicate clogging of underdrain filter mat’l.
–Model overprediction of runoff could also contribute
•Current data do not allow resolution of the issue
2014 Mid-Year GI Data Review: Bioretention
Sites
Recommendations:
•Conduct hydrant tests
•Conduct field observation during rainfall
events, document conditions
•Would help resolve the question of high
capture and could give greater certainty to
performance conclusions
•Without this, reporting on data will need
caveats
2014 Mid-Year GI Data Review
Monroe Rain Garden Site
2014 Mid-Year GI Data Review: Monroe Rain
Garden
Data Summary:
•Continuous radar rainfall
•Outflow monitored continuously: 1/1/14
through 6/30/14
2014 Mid-Year GI Data Review: Monroe Rain
Garden Site
Data Summary:
•Data show good capture
•But, there are some issues with the data
•Overall capture will be less if exclude events
that likely discharged but unable to adjust
outflow data properly
Site # Rainfall/ Runoff
Events
(Jan. 1 – Jun. 30)
Total
Runoff
(ft3)
Total
Outflow
(ft3)
Overall
Capture
Monroe 65 2,600 447 83%
2014 Mid-Year GI Data Review: Monroe Rain
Garden Site
Data Issues:
•Outflow data requires manual adjustment to
avoid losing information due to sub-zero drift
•Data “noise” during some periods
•Data “pulsing” during some periods
•Outflow response larger than runoff
•Outflow response without rain
2014 Mid-Year GI Data Review: Monroe Rain
Garden Site
Data Issues:
•Outflow magnitude > Runoff magnitude
Outflow response
larger than runoff
2014 Mid-Year GI Data Review: Monroe Rain
Garden Site
Equipment
malfunctioning
during coldest
temperatures?
Daily pulsing
2014 Mid-Year GI Data Review: Monroe Rain
Garden Site
Observations:
•One event had higher discharge volume (223
c.f.) than predicted runoff volume (115 c.f.)
•If remove the above event, overall capture
changes from 83% to 91%
•More uncertainty during snow events and
snowmelt events (February 2014)
2014 Mid-Year GI Data Review: Monroe Rain
Garden Site
Recommendations:
•Conduct controlled tests
•Conduct field observation during rainfall
events, document conditions
•Would help resolve the question of high
capture and could give greater certainty to
performance conclusions
•Without this, reporting on data will need
caveats
2014 Mid-Year GI Data Review
Amended Soil Sites
2014 Mid-Year GI Data Review: Labadie and
Warne Amended Soil Sites
Data Summary:
•Continuous radar rainfall at both sites
•Continuous water table level at both sites for
aerated soil areas and compost-amended soil
areas
2014 Mid-Year Review: Labadie and Warne
Amended Soil Sites (Labadie)
0
0.1
0.2
0.3
0.4
0.5
0.6
494.1
494.3
494.5
494.7
494.9
495.1
495.3
495.5
01-Jan-14 21-Jan-14 10-Feb-14 02-Mar-14 22-Mar-14 11-Apr-14 01-May-14 21-May-14 10-Jun-14 30-Jun-14 15-minute Rainfall (in)Elevation (ft)Labadie Amended Soils
Front Back Rainfall (in.)
Water table response
observed during April
2014 Mid-Year Review: Labadie and Warne
Amended Soil Sites (Labadie)
0
0.1
0.2
0.3
0.4
0.5
0.6
494.1
494.3
494.5
494.7
494.9
495.1
495.3
495.5
01-Apr-14 06-Apr-14 11-Apr-14 16-Apr-14 21-Apr-14 26-Apr-14 01-May-14 15-minute Rainfall (in)Elevation (ft)Labadie Amended Soils
Front Back Rainfall (in.)
Water table response evident
in both compost-amended soil
and aerated soil
Slight water table response
evident in compost-amended
soil but not in aerated soil
2013 Data Review: Labadie and Warne
Amended Soil Sites (Labadie)
Event Date 24-hour Rainfall
(in)
Ave. Intensity
During Active
Rainfall (in/hr)
Water Table
Response
Observed?
4/2/14 1.26 0.23 None
4/3/14 1.57 0.11 Compost & aerated
4/13/14 0.94 0.59 Compost only
4/24/14 0.78 0.15 None
2013 Data Review: Labadie and Warne
Amended Soil Sites (Labadie)
Observations - Labadie:
•Response observed in both compost-amended
soil and aerated soil during one (1) rain event
at Labadie
•Two other rain events caused water table
response in compost-amended soils, but not
aerated soils
2014 Mid-Year Review: Labadie and Warne
Amended Soil Sites (Warne)
0
0.1
0.2
0.3
0.4
0.5
0.6
512.4
512.5
512.6
512.7
512.8
512.9
513.0
513.1
1/1/2014 1/21/2014 2/10/2014 3/2/2014 3/22/2014 4/11/2014 5/1/2014 5/21/2014 6/10/2014 6/30/2014 15-minute Rainfall (in)Elevation (ft)Warne Amended Soils
Front Back Rainfall (in.)
Water table response
observed during April
2014 Mid-Year Review: Labadie and Warne
Amended Soil Sites (Warne)
0
0.1
0.2
0.3
0.4
0.5
0.6
512.4
512.5
512.6
512.7
512.8
512.9
513.0
513.1
4/1/2014 4/6/2014 4/11/2014 4/16/2014 4/21/2014 4/26/2014 5/1/2014 15-minute Rainfall (in)Elevation (ft)Warne Amended Soils
Front Back Rainfall (in.)
Water table response evident
in compost-amended soil, but
not in aerated soil
2013 Data Review: Labadie and Warne
Amended Soil Sites (Warne)
Event Date 24-hour Rainfall
(in)
Ave. Intensity
During Active
Rainfall (in/hr)
Water Table
Response
Observed?
4/3/14 1.84 0.12 Compost only
4/7/14 0.38 0.09 None
4/13/14 0.8 0.25 Compost only
4/24/14 0.77 0.13 Compost only
2013 Data Review: Labadie and Warne
Amended Soil Sites (Warne)
Observations - Warne:
•Response observed only in compost-amended
soil during three (3) rain events at Warne
•No response in aerated soils
2013 Data Review: Labadie and Warne
Amended Soil Sites
Conclusions:
•Water table response appears to be
attributable to overall rainfall volume, not
intensity.
•Compost amendment appears to provide
greater infiltration benefit than aeration.
•Response varies between sites: the April 24
event was nearly identical at both sites, but
only Warne showed a response.
2013 Data Review: Labadie and Warne
Amended Soil Sites
Recommendations:
•Continue monitoring
2014 Mid-Year GI Data Review
Humphrey Porous Alley
2014 Mid-Year GI Data Review: Humphrey
Porous Alley
Data Summary:
•Continuous radar rainfall
•Outflow monitored 1/1/14 through 6/30/14,
but several large gaps (of 25, 8, 6, 10, 8 days)
Site # Rainfall/ Runoff
Events
(Jan. 1 – Jun. 30)
Total
Runoff
(ft3)
Total
Outflow
(ft3)
Overall
Capture
Humphrey
Porous Alley
72 153,477 5,798 96%
Humphrey
Porous Alley
(gaps excluded)
30 88,255 5,798 93%
2014 Mid-Year GI Data Review: Humphrey
Porous Alley
Data Issues:
• Sensor drift (upward) on outflow
•Pronounced during May-June time frame
•Requires manual “shifting” of the data
•All events from 1/1/14 through 3/26 and from
4/5 through 5/29 must be removed from data
analysis due to data gaps or data flat-lined at 0
2014 Mid-Year GI Data Review: Humphrey
Porous Alley
Not a data gap,
but only 0
recorded
Sensor Drift
Data Gap
Data Gaps
2014 Mid-Year GI Data Review: Humphrey
Porous Alley
Observations:
•Once manually modified, outflow data show
reasonable responses to rainfall and runoff
amounts
2014 Mid-Year GI Data Review: Humphrey
Porous Alley
Conclusions/Recommendations:
•Peak flow reductions range from 72%-100%
for the 30 events with “good” outflow data
•Volume reductions range from 84%-100%
•Continue monitoring
•Address upward drift if possible through more
frequent site visits
2014 Mid-Year GI Data Review
Summary of Mid-Year Conclusions and
Recommendations
2014 Mid-Year GI Data Review
Summary Conclusions:
•Overall, data collection is greatly improved over
2013
•Project performance (runoff capture) appears to
be very good
•Issues still exist (drift, noise, gaps, inexplicable
results, surprisingly high capture)
•Data processing must be done manually,
requiring more effort than originally expected;
automated data handling not feasible
2014 Mid-Year GI Data Review
Summary Recommendations:
•Conduct hydrant testing as prev. discussed
–Needed to resolve apparent data discrepancies and
strengthen conclusions down the road
•Conduct visual monitoring of projects during
rainfall events
–Will also help us interpret data, make stronger
conclusions
•Perform more frequent instrument maintenance
–Will help avoid long periods of drift, data gaps
2014 Mid-Year GI Data Review
Discussion, next steps...
Semi-Annual Program Review
and
Mid-Course Data Evaluation
MSD Green Infrastructure Pilot Program
February 27, 2015
GI Pilot Program Review
•Semi-Annual Program Review
–Data review by site
•Bioretention sites (Vandeventer, Geraldine)
•Monroe rain garden site
•Humphrey porous alley
–Data summary, issues, observations, conclusions,
recommendations
•Mid-Course Data Evaluation
–Methods
–Results
•Summary, Recommendations, Issues and Next steps
MSD Green Infrastructure Pilot Program
Semi-Annual Program Review
GI Pilot Program Review
Bioretention Sites
GI Pilot Data Review: Bioretention Sites
Data Summary:
•Continuous radar rainfall at both sites
•Underdrain level and outflow monitored
continuously: August 2013 through present
–Several data gaps, especially early in datasets
•Basin level monitored continuously: April 2014
through present
•2 hydrant tests at Vandeventer (10/6/14, 10/8/14)
•1 hydrant test at Geraldine (10/7/14)
Rainfall Depth Event Frequency
Distribution
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.0001 0.001 0.01 0.1 1 10Event Frequency (%)Rainfall Depth (in)
Geraldine Vandeventer Monroe Humphrey TypYear2000
Rainfall Average Rate Event
Frequency Distribution
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.0001 0.001 0.01 0.1 1 10Event Frequency (%)Average Rainfall Rate (in/hr)
Geraldine Vandeventer Monroe Humphrey TypYear2000
GI Pilot Data Review: Rainfall Analysis
Summary/Conclusions
•Maximum single event 1.9-2.0 inches
•Finer resolution of radar rainfall results in
several events between 0.001-0.01 inches
•Otherwise, the events monitored are very
similar to the typical year for depth and
average intensity
GI Pilot Data Review: Bioretention Sites
Hydrant Tests
•Purpose: to resolve apparent data
discrepancies reported during Mid-Year review
•Simple setup: pump a known quantity of water
and observe what happens
GI Pilot Data Review: Bioretention Sites
Hydrant Tests
•Results: Vandeventer (10/6/14 and 10/8/14)
•Inflow = 3455 ft3 (2.5 hr.) and 3955 ft3 (4 hr.)
•Outflow = 907 ft3 and 1150 ft3
•Capture = 74% and 71%
GI Pilot Data Review: Bioretention Sites
Hydrant Tests
•Results: Geraldine (10/7/14)
•Inflow = 1350 ft3 over 2 hours
•Outflow = 0 ft3
•Capture = 100%
GI Pilot Data Review: Bioretention Sites
Hydrant Tests
•Outcomes:
–Better understanding of data and how to correct
–Further investigation of Geraldine (10/24/14)
•Where’s the water going?
–Revise Vandeventer HEC-HMS model
GI Pilot Data Review: Bioretention Sites
Data Correction
–Addressed “data issues” during Mid-Year review
•Noise, sub-zero drift, and outflow without rain
–More knowledge on how to correct, informed by
hydrant test results
–Received maintenance work logs for Geraldine
and Vandeventer
GI Pilot Data Review: Bioretention Sites
(Geraldine)
Maintenance
/ Flushing
GI Pilot Data Review: Bioretention Sites
(Vandeventer)
Data Gap
GI Pilot Data Review: Bioretention Sites
Data Summary:
•Data show good capture
– Excluded 29 events for Vandeventer
– Excluded 10 events for Geraldine
•Similar to Mid-Year results, but more faith
Bioretention Site # Rainfall/ Runoff
Events
(Aug ‘13-Dec ‘14)
Total
Runoff
(ft3)
Total
Outflow
(ft3)
Overall
Capture
Vandeventer 131 109,884 3,067 97.2%
Geraldine 137 46,526 2,377 94.9%
GI Pilot Data Review: Bioretention Sites
Recommendations:
•Continue monitoring Vandeventer and
Geraldine
GI Pilot Program Review
Monroe Rain Garden Site
GI Pilot Data Review: Monroe Rain Garden
Data Summary:
•Continuous radar rainfall
•Outflow monitored continuously: 7/24/13
through 12/31/14
•1 hydrant test (10/17/14)
–Inflow 5-6 gpm for > 2 hours
–Rapid infiltration, no outflow, 100% capture
GI Pilot Data Review: Monroe Rain Garden
GI Pilot Data Review: Monroe Rain Garden
Site
Data Summary:
•Data show good capture
•Data corrections are difficult for this site
Site # Rainfall/ Runoff
Events
(11/1/13-12/31/14)
Total
Runoff
(ft3)
Total
Outflow
(ft3)
Overall
Capture
Monroe 128 5,690 275 95%
GI Pilot Program Review
Humphrey Porous Alley
GI Pilot Data Review: Humphrey Porous Alley
Data Summary:
•Continuous radar rainfall
•Outflow monitored continuously: 8/20/13
through 12/31/14
•Damaged weir Aug 2014 – Dec 2014
GI Pilot Data Review: Humphrey Porous Alley
Data Gap
Data Gap
Not a data gap,
but only 0
recorded
GI Pilot Data Review: Humphrey Porous Alley
Data Summary:
•Data show good capture
Site # Rainfall/ Runoff
Events
(10/31/13-12/31/14)
Total
Runoff
(ft3)
Total
Outflow
(ft3)
Overall
Capture
Humphrey 80 227,708 100,662 51.0%
Humphrey
(excluded after
8/17/14)
52 134,738 7,864 94.2%
GI Pilot Data Review: Humphrey Porous Alley
Observations:
•Outflow data show reasonable responses to
rainfall and runoff amounts
GI Pilot Data Review: Humphrey Porous Alley
Conclusions/Recommendations:
•Peak flow reductions range from 72%-100%
•Volume reductions range from 84%-100%
•Continue monitoring
MSD Green Infrastructure Pilot Program
Mid-Course Data Evaluation
Mid-Course Data Evaluation
Big picture – what are we trying to do?
•Monitoring Protocol says we’ll use a statistical
approach to scaling pilot results to full-scale
–Runoff reduction as dependent variable
–Rainfall parameters as independent variables
(necessary to allow scaling to typical year)
•Need to test this approach (and not at the end
of the pilot program)
Mid-Course Data Evaluation
Key questions:
•Will we have enough data to develop volume
reduction predictions for green infrastructure?
•Can we identify key rainfall variables and do
they make sense?
•What should we do differently for 2015?
•Do we need a major course-correction?
GI Pilot Program Review
Preliminary Multiple Linear Regressions
Preliminary Multiple Linear Regressions
•Preliminary regressions intended to:
–Evaluate to what extent we have usable data for statistical
analysis
–Assess the likelihood of generating regressions that can
be used for predicting volume reductions
–Assess which precipitation variables best predict volume
reduction and evaluate whether those variables make
sense
–Begin to develop strategies for refined analysis following
the completion of data collection
Preliminary Multiple Linear Regressions
•For Humphrey, Geraldine, and Vandeventer sites:
–Percent of volume captured was used as the response variable
–Rainfall characteristics were used as predictor variables
•Total rainfall depth
•Peak rainfall intensity
•Average rainfall intensity
•Total storm duration
•Days since previous rainfall event
–Best subsets method of variable selection was used
–No data transformation on response variable for Humphrey
–Arcsin transformation of response variable for Geraldine and Vandeventer
•Improved normality of residuals
•Common transformation method for percentages
Interpreting MLR Results
•R-squared: measure of goodness of fit of the model
–R-squared = Explained variation / Total variation
•Adjusted R-squared: R-squared value with a penalty for inclusion of additional
predictors
–Useful for selecting preferred regression with variable model sizes
•p-Value: tests the hypothesis that a predictor variable has no effect
–Lower p-Value corresponds to greater confidence that the effect is real
–p-Values less than 0.05 or 0.10 are generally considered statistically significant
•Regression coefficients: mean change in the response variable per unit of
change in the predictor variable
–Less intuitive to interpret if the response or predictor variables have been transformed
•Mallows Cp: can be used in selection of preferred regression
–Lower Mallows Cp is preferred and should approach the number of variables
" D a t a l i m i t e d t o e v e n t s w h e r e m e t e r w a s k n o w n t o b e o p e r a t i n g c o r r e c t l y
( 3 / 2 8 - 6 / 2 9 / 2 0 1 4 )
" V e r y s t r o n g r e l a t i o n s h i p w i t h t o t a l r a i n f a l l d e p t h
" O n e e v e n t c o m b i n e d w i t h e a r l i e r e v e n t
8 2
8 4
8 6
8 8
9 0
9 2
9 4
9 6
9 8
1 0 0
0 . 0 0 0 . 5 0 1 . 0 0 1 . 5 0 2 . 0 0 P e r c e n t V o l u m e R e d u c t i o n ( % ) R a i n f a l l D e p t h ( i n )
G I P i l o t M i d - C o u r s e E v a l u a t i o n : H u m p h r e y
" B e s t s u b s e t s
V a r i a b l e s
R -
s q u a r e d
A d j u s t e d
R - s q u a r e d
M a l l o w s
C p D u r a t i o n
R a i n f a l l
D e p t h ( i n )
A v e r a g e
I n t e n s i t y
( i n / h r )
P e a k
i n t e n s i t y
( i n / h r )
D a y s s i n c e
l a s t e v e n t
1 9 0 . 5 9 0 . 1 9 . 8 X
1 8 0 . 3 7 9 . 6 4 5 . 7 X
2 9 2 . 3 9 1 . 6 5 . 4 X X
2 9 1 . 9 9 1 . 2 6 . 8 X X
3 9 3 9 2 . 1 4 . 8 X X X
3 9 2 . 8 9 1 . 9 5 . 6 X X X
4 9 3 . 4 9 2 . 2 5 . 6 X X X X
4 9 3 9 1 . 8 6 . 7 X X X X
5 9 3 . 8 9 2 . 4 6 X X X X X
G I P i l o t M i d - C o u r s e E v a l u a t i o n : H u m p h r e y
" B e s t r e g r e s s i o n :
V o l u m e R e d u c t i o n ( % ) =
9 9 . 8 3 . 1 0 R a i n f a l l D e p t h ( i n )
1 . 3 2 P e a k R a i n f a l l R a t e ( i n / h r )
+ 0 . 2 8 6 D a y s S i n c e L a s t E v e n t
" p - V a l u e o f R a i n f a l l D e p t h = >