oEPA
United States
Environmental Protection
Agencv
October 2018 | E.PA/600/R-18/365 | www.epa.gov/research
Runoff and Sediment Yield on
the US-Mexico Border, Los
Laureles Canyon
EPA EXTERNAL REPORT
External Report
October 2018
Office of Research arid Development
National Exposure Research Laboratory

-------
Runoff and Sediment Yield on the US-Mexico Border, Los
Laureles Canyon
Trent W. Biggs1, Kris Taniguchi1, Napoleon Gudino-Elizondo2, Eddy Langendoen3, Yongping Yuan4,
Ron Bingner3, Doug Liden5
1 Department of Geography, San Diego State University, San Diego, CA
2 Centra de Investigacion Cientifica y de Educacion Superior de Ensenada, Baja California, Mexico
3 USDA-ARS-National Sedimentation Laboratory, Oxford, MS
4 EPA- Office of Research and Development, Research Triangle Park, NC
5 EPA-Region 9 San Diego Border Office, San Diego, CA
Preface
This document was created to introduce researches completed on the regional applied research effort
(RARE) project "Sediment Load Estimation of the Tijuana River Watershed Under Existing Conditions
and the Future Alternative Scenarios for Best Management Practice Implementation". Excessive
flooding and sedimentation threaten both ecosystems and human populations. On the US-Mexico
border, the Tijuana Estuary in the United States suffers from "excessive sedimentation", which requires
expensive maintenance of sediment traps in the United States. Excessive sedimentation also damages
infrastructure, particularly, in Mexico communities, which result in disruption of services and mortality.
Thus, determining the source of the sediment and mitigating its production is a primary management
goal of the US EPA and other cross-border agencies. Despite the importance of erosion and
sedimentation for the well-being of humans and ecosystems on the US-Mexico border, little data exists
to successfully measure and model the impact of urbanization on watershed processes. This study was
intended to fill this gap by presenting an integrated dataset necessary for supporting comprehensive
study of runoff, soil erosion and sediment production in this region.
Acknowledgements
We acknowledge the contributions of the other agencies including San Diego State University, the
USDA-ARS, Dr. Thomas Kretzschmar, the Centro de Investigacion Cientifica y de Educacion Superior
de Ensenada. Authors would like to thank Dr. Heather Golden, Chi-Hua Huang, Steve Kraemer, and
Roger Kuhnle for reviewing the report and providing helpful feedback and suggestions. Thanks also go
to Fernando Jagueri for help with sediment sampling, and to various residents of Los Laureles for
housing and maintaining equipment.
Disclaimer
The U.S. Environmental Protection Agency through its Office of Research and Development (ORD)
funded and collaborated in the research described herein. The views expressed in this report are those of
the authors, and do not represent and should not be construed to represent any U.S. EPA determination
or policy. Any mention of trade names, products, or services does not imply an endorsement or
recommendation for use. This is a contribution to the EPA ORD Sustainable and Healthy Communities
Research Program.
Contact Information
If you have questions or need additional information, email the project lead Yongping Yuan at
ii

-------
yuan. yongping@epa.gov or professor Trent Biggs at tbiggs@sdsu.edu or go to the project website
(https://biggslab.sdsu. edu/?page_id=50).

-------
Executive Summary
The Tijuana River Watershed (TRW), which drains 4465 km2, including 3,253 km2 (73%) in
Mexico and 1,212 km2 (27%) in US, has experienced large losses of sediment from sheet and rill erosion,
gully formation, and channel erosion. Excessive soil erosion and transport and deposition of sediment in
the watershed have caused many detrimental effects to the people living in the watershed. Communities
on both sides of the US-Mexico border are adversely affected by increased flooding from vegetation
removal and paving, and communities in Mexico in particular experience disruption of services and
transportation due to erosion of unpaved roads. Furthermore, sediment loading from the watershed also
impairs conditions for ecosystems in the Tijuana estuary. This project seeks to address Region 9 Science
Council priorities associated with Sec. 303 [33 USC 1313] and Sec. 319 [33 USC 1329] of the Clean
Water Act (CWA) and the National Environmental Policy Act (NEPA). The long-term research objective
of this project is to gain better understanding on how urbanization affects both ecosystems and human
populations along the US-Mexico border. Realization of the detailed objective to achieve the long-term
protection of ecosystems in the Tijuana estuary and human population includes field data collection and
modeling and analysis of the critical factors impacting different erosion processes (sheet and rill, gully,
channel). The results of this effort will be used to determine an effective approach for sediment loading
estimation as well as evaluating the mitigation of sediment loads that could result from implementation of
conservation easements, re-vegetation, sediment basins, paving, and other Best Management Practices
(BMP).
The first report (Runoff and Sediment Yield on the US-Mexico Border, Los Laureles Canyon)
presents an integrated dataset necessary for supporting comprehensive analysis and modeling of runoff,
soil erosion and sediment production. In order to quantify erosion processes in detail, data collection
focused on Los Laureles Canyon watershed (LLCW), one sub-watershed of the TRW draining to the
Tijuana estuary. The dataset includes rainfall, runoff, suspended sediment concentration, and sediment
yield observed in sediment traps. Secondly, this report describes the rainfall-runoff-sediment relationships
in the watershed and how rainfall type and intensity affect those relationships. Finally, total sediment
yields at the outlet from the LLCW were compared with other natural and urbanized watersheds in
southern California. In this report, section 2.1 describes rainfall data. Section 2.2 describes runoff data,
including depth sensors installed at the outlet, discharge calculation (section 2.2.2), and definition of
events for hydrological analysis (Section 2.2.3). Section 2.3 describes suspended sediment concentrations
and loads, Section 2.4 describes sediment observed in the sediment traps at the outlet and corrections for
trap efficiency, and Section 2.5 describes sediment accumulation in a newly constructed sediment
retention basin in Mexico. Together, the data provide a baseline of water and sediment load across the
border. The central findings include: 1) the rainfall-runoff relationship can be approximated with an SCS
CN of 80-90 which is consistent with the urban land cover; 2) 6-hour rainfall intensity was a key control
on peak runoff production; 3) suspended sediment concentration was relatively stable at high discharge,
with a volume weighted mean of -20 g/L; 4) total annual sediment load observed at the outlet correlated
linearly with annual rainfall, and 5) annual sediment load was -5000 tons km"2 per year, which is ~10x
higher than other urbanized watersheds in southern California and among the highest rates observed in
the southwestern US.
iv

-------
Table of Contents
1. INTRODUCTION	1
1.1.	Background Information and Objectives	1
1.2.	Study Area	1
2. METHODS AND RESULTS	4
2.1 Rainfall	4
2.1.1.	Rain gages in and near the watershed, 2014-2017	4
2.1.2.	Rain gages for sediment traps and modeling	6
2.1.3.	Recurrence inten'als for measured storms	 7
2.2.	Runoff	8
2.2.1.	Pressure transducer and stage correction	9
2.2.1.1	Atmospheric pressure correction	9
2.2.1.2	Observed Stage	9
2.2.2.	Discharge calculation	10
2.2.2.1	Discharge from the PT stage	10
2.2.2.2	Discharge from IBWC bubbler updated rating curve	11
2.2.3.	Event determination	13
2.2.4.	Rainfall-runoff relationships	14
2.3.	Suspended sediment concentrations and event-wise loads	16
2.3.1 SSC measurements and SSC-0 relationships	16
2.3.2.	Event-wise suspended sediment loads (SSL)	18
2.3.3.	Particle size distribution of SSC samples.	19
2.4.	Sediment load in traps at the outlet	19
Trap efficiency and corrected sediment load	20
2.5.	Sediment accumulation in retention basin in Mexico	22
2.5.1.	Sediment sur\>ev after storm in January. 2016	23
2.5.2.	Particle size distribution in the sediment trap in Mexico	23
3. DISCUSSION	25
3.1.	Rainfall and runoff	25
3.2.	Sediment yield at the outlet	26
4. CONCLUSION	27
5. REFERENCES	28
APPENDIX A. HYDROGRAPHS AND HYETOGRAPHS FOR ALL EVENTS	30
APPENDIX B. HYDROGRAPHS DURING SSC MEASUREMENTS	46
APPENDIX C. SEDIMENT TRAP TEXTURE ANALYSES AND TRAP EFFICIENCY	50
APPENDIX D. LINKS TO DATA AND SCRIPTS	58
v

-------
1. INTRODUCTION
1.1. Background Information and Objectives
Excessive flooding and sedimentation threaten both ecosystems and human populations. On the
US-Mexico border, urbanization has increased runoff and sedimentation loads. In the Tijuana-San Diego
region, the Tijuana Estuary in the United States suffers from "excessive sedimentation" (Weis et al, 2001),
and determining the source of the sediment and mitigating its production is a primary management goal
of the US EPA and other cross-border agencies.
Urbanization in Mexico differs from urbanization in the United States, with a longer duration of
the phase of exposed soil (Biggs et al, 2010), where vacant lots and unpaved roads may have high sediment
production for decades. Over a decadal time-scale, urbanized watersheds gradually accumulate
impervious surfaces and earthen channels are lined with concrete, resulting in decreased sediment
production but increased runoff and perhaps increased channel erosion downstream of paved watersheds
and channelized reaches. Communities on both sides of the US-Mexico border are adversely affected by
human-induced watershed alteration, including increased flooding from vegetation removal and paving,
and communities in Mexico in particular experience disruption of services and transportation due to
erosion of unpaved roads. A majority of residents in Los Laureles Canyon on the US-Mexico border report
that they are impacted by flooding and/or road damage due to storms (Grover and Swanson, 2011).
Despite the importance of erosion, sediment loads and runoff for the well-being of humans and
ecosystems on the US-Mexico border, little data exists to successfully measure and model the impact of
urbanization on watershed processes. The objectives of this report are to:
1.	Present an integrated dataset necessary for supporting comprehensive modeling of runoff, soil
erosion and sediment production in a small watershed draining to the Tijuana estuary. The dataset
includes rainfall, runoff, suspended sediment concentration, and sediment yield observed in
sediment traps.
2.	Describe the rainfall-runoff-sediment relationships in the watershed and how rainfall type and
intensity affect those relationships.
3.	Compare total sediment yield at the outlet with other natural and urbanized watersheds in southern
California.
The study included collection of 5-minute rainfall and runoff for ten events (2014-2016), collection
of water samples for measurement of suspended sediment concentration, compilation of data on annual
sediment yield at a sediment trap in the United States (2006-2014) and analyses of sediment texture for
traps in the US and Mexico.
1.2. Study Area
The Los Laureles Canyon watershed (LLCW) lies on the US-Mexico border (Figure 1.1). The
climate is semi-arid Mediterranean, with a wet season from November to April and mean annual
precipitation of-100 mm/yr. The LLCW is on the San Diego Formation, which includes marine and
fluvial sediment deposits that include conglomerate, sandy conglomerate, and siltstone. Soils include
l

-------
cobbly sandy loams and sandy loams, some of which are highly erodible when disturbed. Slopes are
steepest on the incised canyons of the downstream sections of the mainstem. Mean slope as measured by
a 5 m-resolution LiDAR digital elevation model from 2006 is 13.5°, with a maximum slope of 63°.
The watershed drains southeast to northwest. The outlet of the watershed for the purposes of this
study is defined by the outlet of a pair of sediment traps in the United States. The drainage area at the
outlet of the sediment trap is 11.57 km2, with 0.59 km2 in the United States and 10.98 km2 in Mexico
(Figure 1.1).
Mexico
RG.SG
Pacific
Ocean
RG.HM
SW
San Diego
RG.RM
SE
Pac.
Ocean
0.5
2 Kilometers
20 Kilometers
United States
Legend
O Stage Sensor
A Raingage
A NOAA Raingages
|__| Sediment Traps
~ Los Laureles Watershed
|	 Tijuana Estuary
	Channel
Figure 1.1. Rain gage (RG), pressure transducer (PT), field camera, and the US and Mexico sedimentation basin locations with Los
Laureles Canyon watershed boundary. BBLR is the water depth bubbler maintained by the International Boundary Water Commission
(IBWC). Three tributaries are labelled; Main, SW, and SE. The inset location map shows six rain gages near LLCW (white triangles)
maintained by the National Oceanic and Atmospheric Administration (NOAA) or the International Boundary Water Commission
(IBWC).
The watershed was urbanized starting in 1962, with most urbanization occurring between 1980
and 2002 (Figure 1.2). Much of the urbanized area is in "irregular" settlements that are not part of the
formal planning process and are unregulated by the City of Tijuana or other central planning authority.
The watershed is approximately 30% impervious, as calculated with maps from Biggs et al (2010) updated
with visual interpretation of Google Earth™ imagery (K. Taniguchi, unpublished data).
2

-------
In 2005, two sediment traps were constructed in the United States on the course of the main stream
before it flows into the Tijuana Estuary (Figure 1.1). The traps have a combined capacity of approximately
185,804 to 234,830 tons of sediment (based on bulk density of 1.67 tons/m3) and were designed to capture
excessive sediment and prevent it from depositing in the estuary (SWIA, 2001). In 2015, an additional
sediment trap was installed in Mexico, just downstream of the confluence of a major tributary (southeast
channel) but upstream of another tributary (SW channel) (Figure 1.1).
United States
Mexico
Pacific
Ocean
]1 Kilometer
I Open Space
Year Urbanized
1980
I I 1994
I I 2002
Figure 1.2. Map of the urban area by year urbanized (1962-2002) in the Los Laureles Canyon watershed. Data described in
Biggs et al (2010).
3

-------
2. METHODS AND RESULTS
2.1 Rainfall
2.1.1. Rain gages in and near the watershed, 2014-2017
Two rain gages in LLCW collected data for all events that occurred during the study period (2014-
2016), one near the center of the watershed at Hormiguitas (RG.HM, Figure 1.1), and another at the mouth
of the LLCW upstream of the US sediment traps (RG.GC, Figure 1.1). A third rain gage is in a
neighboring watershed 1.2 km east of RG.GC at Smuggler's Gulch (RG.SG, Figure 1.1). A fourth rain
gage was installed in the southeast channel, but was disturbed during subsequent construction at the
installation site and did not record data. A fifth gage was installed at RG.RM in April 2016. A sixth rain
gage installed near RG.RM in 2014 malfunctioned. Rainfall data were collected for all events at RG.HM
and RG.GC and for all events except those between 1/25/2015 and 6/1/2015 at RG.SG; data for that period
were lost during data transfer at the County meteorology office (R. Allen, personal communication, March
2016).
Rainfall at HM was higher than rainfall at GC by an average of 23% and by 17% for total rainfall,
which is consistent with a small orographic effect (Table 2.1). The absolute difference between rainfall
at the low and high elevation sites was consistently between 4 and 9 mm, so the percent difference was
smaller for larger events. Rainfall at GC and SG were very similar for all events, suggesting that rainfall
was homogeneous for a given elevation. The mean elevation of the watershed is 174 m, with -58% of the
watershed area above the elevation of the HM gage and -42% below (Figure 2.1). The HM gage is also
near the centroid of the watershed, so we use it as representative of mean rainfall over the watershed.
The maximum 15-minute, 1-hour and 6-hour intensities were calculated for each event. The
maximum intensity durations were based on the time of concentration (tc) of the watershed, which was
calculated using the Kirpich equation (after Dunne and Leopold, 1978):
_ L115
c ~~ 7700H0-38
where L is the distance from the outlet to the most distant ridge along the main stream (ft), and H
is the difference in elevation between the basin outlet and the most distant ridge (ft). At the outlet of
LLCW, just south of the US-Mexico border culvert, L=22,326 ft and H=616 ft, so tc was 1.09 hr.
4

-------
Table.2.1. Total storm rainfall (mm) at all gages with data. Gage locations are in Figure 1.1.

Elev
m
Total
mm
Date range
Storm #


1
2
3
4
5
6
7
8
9
10
Year


2014
2015
2015
2015
2016
2016
2016
2017
2017
2017
Rain Gage


2/27-3/1
2/28-3/3
5/14-15
9/15-16
1/4-8
3/5-11
4/8-10
1/17-23
2/17-22
2/26-28
RG.GC
36
349
25
27
20
24
44
28
13
59
39
70
RG.SG
40
-
30
-
-
22
46
-
-
58
34
64
RG.HM
174
411
32
36
24
31
50
34
14
65
42
83
RG.RM
187
-
-
-
-
-
-
-
5
-
-
-
SDBF
144
361
30
38
17
22
52
25
19
58
41
59
LIND
5
386
42
26
47
32
76
18
6
48
30
61
IB3.3
40
332
28
35
7
18
53
22
18
47
43
62
HM-GC"
-
38
7
9
4
6
7
5
1
6
3
13
%
-
17
26
35
19
27
15
18
6
10
8
18
HM-SDBF3
-
50
2
-2
7
9
-2
9
-5
7
1
24
%

14
8
-5
41
41
-3
35
-28
13
2
40
HM-LIND3

25
-10
10
-23
-1
-25
15
8
17
12
22
%

6
-24
38
-49
-3
-33
82
131
35
40
36
*. HM-GC, HM-SDBF, and HM-LIND are the difference between rainfall at RG.HM and RG.GC, RG.SDBF, and
RG.LIND respectively
RM
HM
.cP
GC SG
Y
0
50
100
150
200
250
300
Elevation (m)
Figure 2.1. Hypsometric curve of the LLCW watershed showing the elevation of the rain gages.
5

-------
2.1.2. Rain gages for sediment traps and modeling
Rainfall data is also required for interpreting the sediment trap data (2005-2014) and for long-term
simulation of the runoff and sediment yield of the watershed. The gages for this study only collected data
for 2014-2017, and rain gages in Mexico near the LLCW report data only through the 1980s, and so cannot
be used for analysis of the sediment trap data.
Several rain gages are maintained in the US by NO AA or the IBWC (Figure 1.1, inset map), though
the data availability is variable. Only two stations have 100% data coverage for 2005-2012 (SDBF and
Lind) (Figure 2.2), and only one station has 100% coverage for a longer period 1980-2016 (Lind) (Figure
2.3)
SDBF
Jamul
IBream
Figure 2.2. Data availability for rain gages near the LLCW over the period when sediment trap data are available (2005-2016). Y-axis is
the station, and x-axis is the year. Numbers in the grid indicate the fraction of days with data, out of 365.
Jamul
IBream
59.68.68.6®. 70.69.68.68.68.66.69.60.1
Figure 2.3. Data availability for rain gages near the LLCW over 1980-2016. Y-axis is the station name, and x-axis is the year. Numbers in
the grid indicate the fraction of days with data, out of 365.
For the events that have rainfall data in the LLCW watershed at Hormiguitas (RG.HM) (Table
2.1), the gage at San Diego Brownfields (SDBF) has the highest correlation coefficient and smallest
6

-------
RMSE of the stations with good availability (Figure 2.4). Rainfall at RG.HM was higher than that at all
other stations for larger events (>60mm), but matched the SDBF data well for rainfall 10-50 mm (Figure
2.4). The SDBF and HM gages have a higher correlation coefficient and lower error compared with
stations closer to the LLCW in the Tijuana Estuary (IB3.3), so SDBF can be considered to be the best
available option for estimating rainfall in the LLCW for 2005-2016. Selection of the time series for 1980-
2016 for modeling will need to consider the probability distribution of rainfall, which will be analyzed in
a future report.
o
o -
~ SDBF
« Lind




~ IB3.3

^ "

§ ~




o
-------
Table 2.2. Rainfall depth (mm) at different recurrence intervals based on NOAA Atlas 14, Imperial
Beach station, and the observed 15-minute maximum and median rainfall recorded at the mouth of the
LLCW and at the RG.HM station for the 10 monitored storms in Table 2.1.

Duration
NOAA Atlas 14
15 minute
1 hr
6 hr
Recurrence interval (years)



1
4.5
8.7
19.2
2
5.6
10.9
24.1
5
7.1
13.9
30.5
10
8.4
16.3
35.6
RG.HM max
6.0
11.3
33.3

(2014-03-01)
(2015-09-15)
(2017-02-27)
RG. HM median
3.4
6.3
15.1
2.2. Runoff
Water depth was measured at two locations in the main channel near the outlet, upstream of the
sediment traps: 1) on the Mexico side of the border, a pressure transducer (PT) was deployed in the main
concrete channel (Figure 1.1) before each of the 10 rain events in Table 2.1. The PT was housed in a 4"
diameter PVC pipe that extended to the channel bed. In February 2017, a field camera was also deployed
at the site to record water depth every 15 minutes during events, in part due to problems with the stability
of the PT readings observed during large events. The drainage area at the PT is 10.23 km2. 2) On the
U.S. side of the border, the International Boundary Water Commission (IBWC) maintains a bubbler
(Waterlog H350XL pressure sensor and H3551 gas purge system) that records water depth every 15
minutes and transmits data to San Diego County ALERT system (https://sandiego.onerain.com/). The
IBWC bubbler is not in an ideal location for discharge measurement, since it is located in a concrete
reservoir to the side of the main channel that is separated from the main flow by a set of poles that intercept
debris (Figure 2.5), creating possible hysteresis in the stage-discharge relationship. Here, we developed
a rating curve for the IBWC gage for a large storm event in February 2017 that had data from the field
camera installed in Mexico (Section 2.2.2.2). We then used that rating curve to estimate discharge at the
IBWC gage for all events, and compared the IBWC discharge with the discharge recorded by the PT.
After events were defined, the rainfall for each event was compared with the Storm Types defined by the
Soil Conservation Service (SCS) Technical Report 55 (SCS, 1986). Types I and IA are common in Pacific
Maritime climates including California, Type III are typical of tropical storms in Atlantic coastal areas
and Gulf of Mexico, and Type II is most common in the rest of the continental United States, and has the
highest short-duration (1-6 hr) rainfall intensities.
8
-------
Figure 2.5. IBWC bubbler located on the side reservoir of the Main channel. Flow from the Main channel passes through the US-
Mexico border culverts (A) into the Tijuana Estuary. Photo credit: K. Taniguchi, February 2017.
2.2.1. Pressure transducer and stage correction
2.2.1.1	Atmospheric pressure correction
The PT at the outlet in Mexico measured total pressure (water + atmosphere). In order to calculate
water depth, atmospheric pressure from a nearby weather station is subtracted from observed pressure at
the PT. For the sampled events, atmospheric pressure data were taken from a nearby barometer, Tijuana
Estuary Naval Auxiliary landing field (TJE NAVAL), station USAF 722909, NCDC 93115 in Imperial
Beach, CA (https://gis.ncdc.noaa.gov/maps/ncei/cdo/hourlv). The station is ~4 km from the outlet of
LLCW, but is at the same elevation. Atmospheric pressure at TJE NAVAL and two other nearby
barometric stations was higher than total pressure observed by the PTs, though the offset was constant
over short (hourly) time intervals (Figure 2.6). For each storm, several offsets were calculated by
assuming that water depth was zero after extended periods (~4 to 6 hr) of no rainfall, and the offset was
subtracted from the atmospheric pressure time series (Figure 2.6B). The final water level was calculated
as the difference between PT total pressure and the adjusted atmospheric pressure at the TJE Naval Base
(Figure 2.6). Any negative stage values were replaced with zero.
2.2.1.2	Observed Stage
A staff gage was painted on the side of the channel in January 2016 for comparison with stage
calculated from the PTs and atmospheric pressure. Only a few observations were available in 2016 and
2017. On 2016-01-06, video of channel flow was taken at 11:00 am by residents living at the gage
location. Based on markings on the channel side, the stage at this point in the January 6, 2016 event was
60 cm, which compares well with the peak stage reported by the pressure transducer (57 cm). See link
below for video: https://www.voutube.com/watch?v=lRABWKisSDE&feature=:voutu.be
On 2016-03-06 at 9:40 am, the maximum stage verbally reported by residents was 25 cm,
compared with a maximum stage recorded by the PT of 6 cm. Stage was also measured with a camera
and visually in January and February 2017, but the PT failed so no comparison was possible. Photographs
of the channel were taken during water sampling activities on 2014-12-03, 2014-12-12, 2014-12-13, 2014-
12-13, 2014-12-17, 2015-03-01, 2015-05-14, and 2017-01-21. The photo taken on 2015-05-14 shows
9
-------
flow in the channel when the PT recorded zero stage, due to a slightly uneven concrete channel bottom
that focused flow in part of the channel not in contact with the PT.
More data is needed to validate the stage recorded by the PT. The limited visual stage observations
suggested that the PT underestimated water depth by several cm, but insufficient data were available to
develop a single correction factor. The proportional error should be smaller for peak discharge assuming
that the underestimation is a constant offset of a few cm.
See Appendix A for a complete description of all observed storm events.



CO
F

b

s=
CN "
01

ir

F
O _
>

o
IT) -

o -
A)
E1.PT
E2.IBWC
E3JBWC

o

CN

o '

to

"r~
F
o


0)





(D
o

o

o '


O)
-

CN -

< >

-------
different times from floating debris, which was used to back-calculate a Manning's n ranging from 0.008
to 0.012. Based on these field observations and the Manning's n for "ordinary concrete lining" (0.013,
Dunne and Leopold, 1978), the Manning's n used in all calculations of discharge was 0.013. Sediment
accumulated in the reach with the PT after some events, which would increase the Manning's n, thereby
decreasing the estimate of discharge. Increasing Manning's n from 0.013 to 0.017 ("concrete and earth
channels in best condition") would decrease the calculated discharge by 24%.
2.2.2.2 Discharge from IBWC bubbler updated rating curve
The IBWC developed a stage-discharge rating curve for the IBWC bubbler using HEC-RAS (S.
Smullen, unpublished data). An updated rating curve was developed for the IBWC bubbler using
discharge calculated from the field camera, using Manning's equation and roughness of 0.013, and field
measurements of flow velocity for a large storm event in February 2017. Separate rating curves were
developed for the rising and falling limbs of the hydrograph, due to potential hysteresis caused by the side
reservoir where the bubbler is located (Figure 2.7). The following rating curve equations were developed
for the rising and falling limbs:
n _ 0	s < 0.91 m
Arising ~ 19£s _ 1?J6 s > Qm m	V)
n _ 0	s < 0.98 m
Vfalling ~ 212s _ 20_y7 s > Q_98 m	( )
where 5 is the stage at the IBWC BBLR (Figure 1.1). The IBWC BBLR provided reliable estimates
of discharge only for the largest events, or for smaller events that were preceded by another event that
filled the small reservoir that houses the IBWC bubbler. For example, see Figure A9, Storm #6, when the
first event (6.5 mm) produced runoff at the PT but an insufficient increase in water level at the IBCW
bubbler to result in a non-zero estimate of discharge, while the subsequent event (23 mm rainfall) produced
a rise in water level sufficient to result in an estimation of discharge from the PT-IBWC rating curve (Eq
2 and 3). The PT appeared to record reliable stages during small events, but fluctuated erratically for
larger events, so for larger events, the IBWC gage and PT-IBWC rating curve were used. See Table 2.3
for which stage reading was used to calculate discharge for each event, and Appendix A for detailed
comparison of the PT and IBWC data.
11
-------
Table 2.3 Summary of storm events defined in Table Al, final observed data (obs) used for calibration/validation.
TCM = thousand cubic meters.
Event Date*
Rainfall
(mm)
Peak Discharge
(cms)
Total Runoff
Runoff Ratio
(Q/P)

Event
Source


Obs
Obs
(mm)
Obs
(TCM)
Obs



Storm 1

2014-02-28

12.25
1.13
0.27
2.80
0.02

El
PT

2014-03-01

7.50
1.54
0.33
3.36
0.04

E2
IBWC

2014-03-02

7.50
6.14
1.08
11.04
0.14

E3
IBWC
Storm 2

2015-03-01

23.25
3.36
1.36
13.88
0.06

El
PT

2015-03-02

9.25
1.43
0.48
4.86
0.05

E2
PT
Storm 3

2015-05-15

22.50
19.46
5.93
60.63
0.26

El
PT
Storm 4

2015-09-15

30.75
5.27
6.40
65.50
0.21

El
PT
Storm 5

2016-01-05

22.25
17.72
3.76
38.42
0.17

El
PT
Storm 6

2016-03-06

6.50
1.03
0.93
9.47
0.14

El
PT

2016-03-07

23.00
5.07
4.23
43.32
0.18

E2
IBWC
Storm 8

2017-01-19

13.00
5.37
2.57
26.27
0.20

El
IBWC

2017-01-20

28.00
6.86
18.66
190.95
0.67

E2
IBWC
Storm 9

2017-02-17

33.25
11.16
7.03
71.89
0.21

El
IBWC
Storm 10

2017-02-27

83.00
16.69
42.07
430.50
0.51

El
CAMERA
*For storm 7, no PT data were available and IBWC rating curve discharge was zero.
Future work could attempt to extend the PT-IBWC rating curve for lower discharges, though the
rating curve for low discharges will be complicated by hysteresis as the reservoir containing the IBWC
bubbler fills during and empties after events. An alternative would be to install a bubbler closer to the
outlet of the channel, on the channel side of the debris barriers (to the right of the yellow debris catch
poles in Figure 2.5).
12
-------
to
E
o
-------
and the third event was set to 2014-03-02 (Table 2.3). The fourth event was small and excluded from the
model calibration and validation. Event 1 did not fit any storm type as defined by the Technical Release
55 (TR-55: Soil Conservation Service, 1986), but was adequately modeled by storm Type 2, 12 hour
(Figure 2.8).
Storm Type comparison March2014
Figure 2.8. Cumulative rainfall fraction, normalized to storm total rainfall, for the two events in March 2014. The four groups of three
lines correspond to the 6-hour (left-most group), 12-hour, 18-hour, and 24-hour design events from TR-55 (Soil Conservation Service,
1986).
2.2.4. Rainfall-runoff relationships
Event rainfall for the 14 events ranged from 7 to 83 mm. The 6-hour rainfall intensity ranged from
1.5 to 33.3 mm. The 1-hour intensity ranged from 0.5 to 11.3 mm, and the 15-minute intensity ranged
from 0.25 to 6.0 mm.
Event total runoff increased with event-total rainfall (Figure 2.9). The runoff coefficients (Q:P)
ranged from 0.02 to 0.67 (Table 2.3). The largest event (rainfall 83 mm) had a runoff coefficient of 0.51.
Most events fell between SCS Curve Numbers (CN) 80 and 90 (Figure 2.9), which is consistent with the
urban land cover in the watershed. The CN was highest for the smallest events and generally decreased
with event size. This is consistent with runoff production from surfaces with low infiltration capacity
during small events, and from all surfaces, including those with high infiltration capacities, during large
events.
14
-------
CN=90
CN=90
Event Rainfall (mm)
Figure 2.9. Rainfall-runoff relationship for all observed storm events summarized in Table 2.3, with several SCS CN rainfall-runoff
relationships, in non-log (top) and log-log (bottom).
Peak event discharge (Qpk) was predicted better by the event-maximum 6-hour rainfall intensity
(Pearson r 0.71, p<0.01) than by the total event rainfall (Pearson r 0.58, p<0.05) (Figure 2.10), mostly due
to two high-Qpk outliers. Qpk was not predicted well by the 15 minute (p>0.1) or 1 hour (p>0.05)
maximum intensity rainfall.
15
-------
o
CM
in
in —
A. Event total
o
CM
m —
20
40
60
80
B. 6-hr maximum
15	20
Rainfall, mm
Figure 2.10. Peak event discharge (Qpk) versus: A. event total precipitation and B. 6 hour maximum precipitation for the events in
Table 2.3.
2.3. Suspended sediment concentrations and event-wise loads
2.3.1 SSC measurements and SSC-Q relationships
Water samples were collected using grab sampling during storm events at the outlet of the
watershed. Due to the high flow velocities and/or small water depths at the time of sampling (often < 30
cm), it was not possible to use a depth-integrated sampler, so a surface grab sample was taken. The water
samples were filtered using pre-weighed filters with nominal pore size of 0.45 |im that were then dried
and reweighed to calculate the suspended sediment concentration (g L"1). Discharge estimated at the PT
was zero for some SSC sampling times due to the position of the PT on the side of the channel. Based on
photographs taken during the SSC samples, an estimated minimum discharge (Qmin) at the time of SSC
sample collection was 0.07 m3/s, which corresponds to 2 cm of flow depth. All discharge values less than
0.07 m3/s at the time of SSC sampling were assumed to be equal to Qmin.
The relationship between SSC and Q was highly variable at low discharges (<0.1 m3 s"1) (Table
2.4, Figure 2.11). SSC during relatively low discharges was high (> 18 g L"1) and was only slightly lower
than the maximum observed SSC (27 g L"1). The variation in the SSC-Q relationship could not be ascribed
to hysteresis in the SSC-Q relationship because no event included samples on both the rising and falling
16
-------
limbs of the hydrograph, and because low concentrations were observed both before and after the peak
(Appendix B).
Table 2.4. Suspended sediment concentration (SSC) for all collected samples.
Storm
Date
SSC (g L"1)
Q (m3 s1)
Storm 1
2014-02-28 17:20:00
3.23
1.54

2014-03-01 07:40:00
0.35
0.07*

2015-03-01 07:20:00
15.20
0.33
Storm 2
2015-03-01 12:40:00
19.40
0.07*

2015-03-01 17:45:00
13.70
0.07*

2015-03-02 08:40:00
3.20
0.07*
Storm 3
2015-05-15 08:30:00
15.85
0.14

2016-03-06 09:40:00
16.34
0.42

2016-03-06 12:00:00
14.58
0.39
Storm 6
2016-03-06 15:45:00
18.53
0.11

2016-03-07 06:00:00
15.21
0.07*

2016-03-07 07:00:00
16.37
0.07*

2016-03-07 12:00:00
11.70
0.07*

2017-02-17 16:00:00
0.22
0.07*
Storm 9
2017-02-17 17:00:00
0.23
0.07*

2017-02-17 20:30:00
20.01
4.11

2017-02-27 12:37:00
15.56
9.91

2017-02-27 13:13:00
21.93
5.52
Storm 10
2017-02-27 13:34:00
27.02
6.18

2017-02-27 14:00:00
24.13
4.00

2017-02-27 14:05:00
10.98
4.00

2017-02-27 14:30:00
20.83
7.97

2017-02-27 15:00:00
14.44
6.18

2017-02-27 15:35:00
24.81
2.67

2017-02-27 16:02:00
11.51
2.67
* Recorded Q was <0.07 m3 s"1 (stage of <0.02 m); assumed minimum water depth of 0.02 m and
Q of 0.07 m3 s"1.
The exponent on the relationship between Q and SSC (0.32) was lower than observed in most other
watersheds in the literature, where b ranges from 0.38 to 2.0 (Syvitski et al, 2000). Relatively flat rating
curves as observed in the LLCW are indicative of highly erodible material that can be transported for a
wide range of flows (Asselman, 2000).
17
-------
CD
CJ
CO
CO
CO
J*
CO
CO
0.5
0.2
100
0.1
0.2
0.5
1.0
2.0
5.0
10.0
Q, m3 s"1
Figure 2.11. Relationship between discharge (Q) and suspended sediment concentration (SSC) and suspended sediment load for the
samples collected in the LLCW. The dashed line is the linear regression fit without bias correction, and the dotted line in the Q-SSL plot
is with bias correction.
2.3.2. Event-wise suspended sediment loads (SSL)
The suspended sediment load (SSL) was calculated for each event with SSC data using four
methods: 1) as the product of the event total discharge and the volume-weighted-mean (VWM) of SSC of
the individual grab samples for that storm, 2) as the product of the event total discharge by the VWM SSC
for all samples from all storms, 3) using the Q-SSL rating curve (Figure 2.11) to estimate SSL for each
15-minute discharge value during the storm, and 4) using the same rating curve approach as in 3), but with
the bias correction factor (bcf) as described by Crawford (1991), which is based on the suggestion of Duan
(1983). The bcf corrects for the underestimation of SSC the results from the use of ordinary-least squares
on log-transformed data.
The SSL varied over two orders of magnitude for the observed storms (Table 2.5). SLL varies by
a factor of ~2 depending on the method used to calculate it, with a factor of 5.6 difference for one storm
on 2014-02-28.
18
-------
Table 2.5. Total event suspended sediment concentration (SSC) and load (SSL) at the PT location for the events with
SSC data.

Event

N SSC

Total Q

ssca

SSL (tons)





Mm
TCM

gL"1

Event
VWMb
All
VWM°
Rating, no
bcf1
Rating,
bcf

2014-02-28 E2

1

0.3
3.4

3.2

11
62
35
56

2015-03-01 El

3

1.3
13.8

15.2

209
255
151
246

2015-05-15 El

1

5.9
60.5

15.8

958
1118
1408
2289

2016-03-06 El

3

0.9
9.1

15.9

144
168
78
126

2017-02-17 El

3

7.0
71.9

20.0

1438
1329
1356
2204

2017-02-27 El

9

42.1
430.6

19.0

8199
7961
8657
14073
a Volume-weighted mean suspended sediment concentration.
b = Q x VWM (volume weighted mean SSC concentration) for samples collected during the event.
0 = Q x VWM (volume weighted mean SSC concentration) for all samples, all events.
d Calculated from the Q-SSL rating curve, with no bias correction factor.
e Calculated from the Q-SSL rating curve, with bias correction factor.
2.3.3. Particle size distribution of SSC samples
The particle size of three SSC samples from April 2016 were analyzed on a laser particle size
analyzer. The time was not recorded for these samples by the in-fteld volunteer collectors, so their
relationship to the hydrograph is unknown. They have a large silt percentage (70-80%, Table 2.6).
The particle size from this and other events will be compared to the texture of soil in the watershed
in future reports.
Table 2.6. Particle size of SSC samples collected in April 2016. Time and discharge for the collection is unknown.
Type
Diameter (|im)
Sample 1 (%)
Sample 2 (%)
Sample 3 (%)
Medium sand
250-500
0
0
0.2
Fine sand
125-250
2.6
0.8
4
Very fine sand
63-125
10.5
5.5
10.7
Silt
4-63
73.5
79.4
75.7
Clay
<4
13.4
14.3
9.4
Median (^m)
10.5
5.5
9.4
2.4. Sediment load in traps at the outlet
The two sediment traps in the United States (Figure 1.1) were completed in late 2004. Data on
sediment removed from the traps were available from the Tijuana River National Estuarine Research
Reserve (TRNERR). Both upper and lower traps were cleaned out in spring and fall 2005, winter 2006,
and each fall from 2007-2012. Starting in 2013, the lower trap was not excavated due to low rainfall
19
-------
(Table 2.7). Topographic surveys were conducted in Fall 2011 (both upper and lower traps) and Fall 2015
(upper trap only).
Table 2.7. Time series of sediment removed from LLCW (Goat Canyon) traps. Data from Chris Peregrin and Cara
Stafford at TRNERR. Mass is calculated using bulk density of 1.67 tons nr3. TL=based on truck loads, CE=complete
excavation. UP=upper basin only excavated. PE=partial excavation NE =no excavation either basin.
Removal date
Volume
removed
(yd-3)
Mass removed (tons)
Trap
efficiency
Notes


Uncorrected
Corrected


2005-03
55,000
-
-

TL, CE
2005-10
35,000
-
-

TL, CE
2006-12
25,000
31920
34642
0.92
TL, CE
2007-10 or-11
25,000
31920
33079
0.96
TL, CE
2008-09
40,000
51072
64580
0.79
TL, CE
2009-10 or-11
45,400
57967
68949
0.84
TL, CE
2010-09 or-10
55,000
70224
78935
0.89
TL, CE
2011-09
50,733
64776
70965
0.91
URS/NV5 survey, CE
2012-09 or-10
45,000
57456
58513
0.98
TL, CE
2013-09 or-10
14,967
-
-

UP, PE
2014-09 or-10
0
-
-

NE
2015-09 or-10
17,963
-
-

Rick Engineering survey: UP, PE
Mean

52190
58523
0.89

Trap efficiency and corrected sediment load
The total sediment yield includes sediment retained in the trap and sediment that flowed through
the trap and entered the estuary. The trap efficiency, which is the proportion of the total sediment yield
that is retained in the sediment basin, was calculated based using Urbonas and Stahre (1993), which is for
turbulent and non-ideal conditions:
E = 1-
1 o.)
n 0)
-n
(4)
where E (range 0 to 1) is the trap efficiency of the sediment in the size class corresponding to
particle fall velocity co; coc is the critical velocity of the basin, which is the fall velocity of the smallest
particles that are 100% retained; and n is a factor that depends on the hydraulic efficiency of the basin. A
range of n values (n = 1 and n = 3) was used in calculating trap efficiency, where n = 1 represents poor
20
-------
settling conditions and n = 3 represents good settling conditions, both for turbulent and non-ideal
conditions (Morris and Fan, 1998). Turbulent and non-ideal conditions were used to give a lower-bound
estimate of the trap efficiency. Methods for calculating co and coc are in Appendix C.
The trap efficiency varies by particle size, storm event and year. A mass-weighted annual trap
efficiency was calculated for each year and particle size as:
Eann = (ZQiEd/lQi	(5)
where Qi is the mean daily discharge on day i and E, is trap efficiency on day i. Daily Q was
estimated from a coupled AnnAGNPS-CONCEPTS model (in preparation). The resulting Eann by size
class allowed for correction of the observed sediment loads to a total sediment load by size class.
Corrected load was calculated by using n=3 for Eann, representing good settling conditions. The particle
distribution was taken from that observed in the upper sediment trap (AMEC, 2007) (Appendix C). More
than half of the sediment in the traps is sand (Table CI), and the median grain size is fine sand. There
was no statistical difference in median particle size between the upper and lower basins, so the upper basin
was used to calculate the trap efficiency, de Temple et al. (1999) reported somewhat more sand in surface
samples of the estuary near the LLCW outlet (Table CI).
The annual trap efficiency varied from 0.79 to 0.98, and was 0.89 for the cumulative mass removed
over 2006-2012 (Table 2.7). Details of the trapping efficiency by particle size are in Appendix C (Table
C2).
Total sediment accumulation in the traps correlates with precipitation at both Lindbergh Field
(Lind) and San Diego Brownfields stations (Figure 2.12). The relationship is linear, which is unexpected
given the usually non-linear relationship between rainfall and sediment load (Inman and Jenkins, 1999).
21
-------
sz
o
SH-
OT
80 -
60 -
40 -
20 -
0 -
A. Lind
"O
OT
O
"E
~C
(f)
80
60
40
B. SDBF

20
0
• Uncorrected
Corrected
100
200
300
400
500
600
Precipitation between cleanings, mm
Figure 2.12. Sediment load to the LLCW (Goat Canyon) sediment traps traps versus total precipitation between cleanings, 2005-2012.
The uncorrected values (black) are the tons of sediment removed from the trap between cleanings, and the corrected values (grey) are
calculated using the retention efficiency for each particle size class (sand, silt, clay). Precipitation is from A. Lindbergh and B. San
Diego Brownfields station.
2.5. Sediment accumulation in retention basin in Mexico
A retention basin was installed in the main channel of LLCW during the project period (2012-
2014, Figure 2.13) downstream of the confluence of the main and southeast channel, but upstream of the
confluence of the main and southwest channel (Figure 1.1). The basin dimensions are approximately 25-
33 m wide, 5-6 m deep, and 172 m long. Based on Google Earth imagery, construction began in November
2012 (outlet structure was built), concrete was poured in winter 2013-14, and the project was finalized by
July 2014.
22
-------
Figure 2.13. Soil depth survey in the Mexico sediment basin in LLCW, taken on January 15, 2016. The sedimentation basin was
installed in winter 2013-14. Yellow pins indicate locations of depth measurements (Table 2.8) and blue drop symbols indicate the
locations of soil sample collection for particle size analysis (Table 2.9).
2.5.1.	Sediment survey after storm in January, 2016
Following the storm from January 4-8, 2016 (total rainfall 49.8 mm), sediment accumulated in the
new retention basin. A survey of sediment depth was conducted after the storm (Figure 2.13, Table 2.8).
Sediment depth was fairly uniform in the basin, ranging from 33-64 cm, with a mean depth of 55 cm. The
total sediment accumulation during the event was approximately 3,600 tons. The total sediment yield at
the outlet of the watershed at the US-Mexico border estimated from the VWM for all samples as in Section
3.2 is 38.42 thousand m3 x 18.5 g/L = 711 tons, or just 20% of what was retained in the Mexico trap. This
could indicate that either the sediment retained in the trap represented a significant fraction of the total
watershed load, or that the load estimated from the VWM is an underestimate. Future sampling and
comparison with sediment accumulation in the traps in Mexico and the US will help substantiate the
sediment budget and the impact of the Mexico trap.
2.5.2.	Particle size distribution in the sediment trap in Mexico
Samples of sediment that accumulated in the retention basin were collected on January 15, 2016
and analyzed for particle size distribution on a laser particle analyzer at UABC, Ensenada. The samples
were dominantly very fine to fine sand, with very little clay (Table 2.9, Figures 2.14 and 2.15).
23
-------
Table 2.8. Sediment depth at the new retention basin on
January 15, 2016. IDs correspond withFigure 2.13.
ID
Sediment depth (cm)
1
33
2
55
3
Cobble
4
63
5
55
6
64
7
62
Mean
55
Table 2.9. Particle size of sediment samples in the sediment trap in Mexico, January
2016. All samples taken from 2-10 cm depth. Samples had no gravel or cobble. Location
codes correspond to Figure 2.13.
Location code
Lab
code
Photo
Figure
Particle size percentage (%)



Sand
0.063-2mm
Silt
4-63 um
Clay
<4um
8.1
9
C4
79.2
18
2.8
8.2
10
-
74.2
23
2.8
8.3
11
-
78.2
19
2.8
15.0r
100

10.0	100
Particle size / (jm
Figure 2.14. Particle size distribution at the sediment basin (location code 8.1)
24
-------
15.0,
10.0	100
Particle size / |jm
Figure 2.15. Particle size distribution at the sediment basin (location code 8.3). Silt is 3.9-62.5 um on the phi scale. Very fine sand is
62.5-125 and fine sand is 125-250 um.
3. DISCUSSION
3.1. Rainfall and runoff
The rainfall events fit a mix of type I (N=6) and type II (N=7) storms (Appendix A), as defined by
SCS (1986). Type I storms are typical of Pacific Maritime climates, with lower storm intensities over
short durations. Type II events are typical of the rest of the continental United States outside the Gulf
Coast, and have the highest short-duration intensities. Our data suggest that both high-intensity type II
and low-intensity Type I storms can occur in the study area, with some storms showing short durations
and high intensities (e.g. 6T2). The relationship between storm characteristics and key driving
mechanisms and moisture sources is not determined here but would be helpful for future research, since
storm type can influence key attributes of watershed response, including storm runoff, peak discharge,
and sediment generation.
The event-total rainfall and runoff for the observed events are consistent with an SCS CN of
between 80 and 90, with decreasing CN for larger events (Table 3.1). The hydrologic soil group of the
LLCW is assumed to be type B based on the soil type (cobbly sandy loam and sandy loam), but in places
may be a type C that has an impeding layer. The CN range for the observed events in the LLCW is
consistent with the CN for urban land use with an impervious cover of between 30 and 65%, compared
with impervious cover in the LLCW of-30%. Soil moisture is critical for runoff production in semi-arid
watersheds, so more detailed modeling that accounts for soil moisture impacts on runoff production may
explain variation in runoff among events.
25
-------
Table 3.1. SCS Curve numbers (CN) from the literature (Dunne and Leopold, 1978 Table 10-8) and for the
LLCW.

Hydrologic Soil Group

B
Moderate infiltration rate, moderately
deep to deep, moderately fine to
moderately coarse textures
C
Slow infiltration rates when
wetted; often have impeding
layer, or moderately fine to
fine texture
Residential, 65% impervious
85
90
Residential, 38% impervious
75
83
Residential, 30% impervious
72
81
Dirt road
82
87
LLCW, -30% impervious
-80-90
3.2. Sediment yield at the outlet
The observed sediment yield in the LLCW (4-5 kt km"2 y"1) was higher than almost all measured
yields from small watersheds in California (Table 3.2), though the sediment yield from undisturbed
chaparral land cover on erodible sedimentary formations in the Western Transverse Range can be as high
as 5.3 kt km"2 y"1 (Warrick and Mertes 2009). An urbanized watershed in southern California (drainage
area 288 km2) with severe channel erosion yielded 0.5 kt km"2 y"1 (Trimble 1997), which is 10% of the
yield from LLCW, though Trimble (1997) did not report gully formation on hillslopes, which was a major
process generating sediment in the LLCW.
The high sediment yield from the LLCW was due in part to the high urban cover percentage (86%)
that has a high (30-40%) bare soil cover fraction (Biggs et al, 2010), including construction sites and
unpaved roads that showed signs of severe erosion, including rills and gullies. A global survey (Russell
et al, 2017) shows that construction sites have between 21 and 11,613 times the sediment yield of the
undisturbed background, and that urban areas can have 1.7 to 68 times the sediment yield as the
background. Unpaved roads also have very high sediment yields, including 125 kt km"2 y"1 for heavily-
used logging roads in the Pacific Northwest (annual rainfall 390 cm) (Reid and Dunne 1984) and 11 kt
km"2 y"1 for recently graded roads in the US Virgin Island (annual rainfall 115 cm). Given that unpaved
roads in the LLCW showed signs of severe erosion similar to what was observed on construction sites,
the observed range of 4.5-5.0 kt km"2 y"1 is expected from a watershed that has a large fraction of its
surface in a condition similar to a construction site.
26
-------
Table 3.2. Sediment yield from watersheds in California compared with yield from LLCW.
Location
Watershed
area km2
Rainfall
mm yr1
Sediment Yield
tons knr2 vr1
Land cover
Reference
Southern CA
118 -10,760
250-650
20-4200
Mixed natural, ag., urban
Inman and Jenkins
1999
Transverse Range, CA
9 -1,000
397-877
600-25003
Natural, some habitation
Scott, 1968
Transverse Range, CA
14-4,185
400-700
740-5300
Natural vegetation
Warrick and
Mertes, 2009
Southern CA
288
330
500
Urban
Trimble 1997
Los Laureles Canyon
11.6
100-330
4499-5040
Urban
This study
a. Reported in m3 km 2, converted using 1.67 tons m~3
4. CONCLUSION
The data collected allowed us to perform a comprehensive assessment of rainfall, runoff, and
sediment load in the LLCW on the US-Mexico border. The observations suggest that:
1.	Rainfall intensity is a critical control on event peak discharge, and event rainfall-runoff
relationships are consistent with a SCS CN that is consistent with a partially urbanized watershed
(CN 80-90).
2.	Event-mean suspended sediment concentration (SSC) was relatively stable for a wide range of
discharge, up to a maximum of 27 g/L. The slope of the Q-SSC relationship is low, indicating that
sediment in the watershed is highly mobile sediment and is transported at a wide range of flows.
3.	Annual total sediment load observed in sediment traps at the outlet correlates linearly with annual
total rainfall. This is somewhat unexpected given the non-linear relationship between rainfall and
runoff, and given previous observations in semi-arid regions (Inman and Jenkins, 1999), and may
be due to overestimation of the trap efficiency for higher annual loads.
4.	Annual sediment yield is higher than in most other watersheds in California, and is consistent with
extremely high rates of erosion.
Future work will quantify the roles of different erosion processes in the sediment budget, and will
model the production of sediment under different land cover and management scenarios.
27
-------
5. REFERENCES
AMEC Earth & Environmental Inc. (2007). Goat Canyon Retention Basin Soil Particle Size Distribution Study.
San Diego, CA.
Asselman, N. E. M. (2000). Fitting and interpretation of sediment rating curves. Journal of Hydrology, 234(3-
4), 228-248. doi.org/10.1016/S0022-1694(00)00253-5.
Biggs, T. W., Atkinson, E., Powell, R., & Ojeda-Revah, L. (2010). Land cover following rapid urbanization on
the US-Mexico border: Implications for conceptual models of urban watershed processes. Landscape and
Urban Planning, 96(2), 78-87. doi.org/10.1016/j.landurbplan.2010.02.005
Crawford, C. G. (1991). Estimation of suspended-sediment rating curves and mean suspended-sediment loads.
Journal of Hydrology, 129(1-4), 331-348.
de Temple, B., Ruttenberg, D., Perala, N. C., & Battalio, R. (1999). A preliminary design of a sediment retention
facilities for Goat Canyon Creek. Corte Madera, CA: Phil Williams and Associates, Ltd.
Dunne, T., & Leopold, L. B. (1978). Water in Environmental Planning. New York: W.H. Freeman and
Company.
Grover, R., & Swanson, K. (2011). Local perspectives on environmental degradation and community
infrastructure in Los Laureles Canyon, Tijuana, Mexico. Geography. San Diego State University, San
Diego.
Inman, D. L., & Jenkins, S. A. (1999). Climate Change and the Episodicity of Sediment Flux of Small California
Rivers. The Journal of Geology, 107, 251-270.
Morris, G.L. and Fan, J. (1998). Reservoir Sedimentation Handbook: Design and Management of Dams,
Reservoirs and Watersheds for Sustainable Use. New York: McGraw-Hill Companies, Inc.
NOAA (National Oceanic and Atmospheric Administration) (2017). Precipitation Frequency Data Server, url:
http://hdsc.nws.noaa.gov/hdsc/pfdsA access date 2017-06-07.
Rubey, W.R., 1933. Settling velocities of gravel, sand and silt particles. American Journal of Science, 21, 325-
338.
Russell, K. L., Vietz, G. J., & Fletcher, T. D. (2017). Global sediment yields from urban and urbanizing
watersheds.	Earth-Science	Reviews,	/^'(Supplement	C),	73-80.
http: //doi. org/https: //doi. org/10.1016/j. earscirev.2017.04.001.
Soil Conservation Service. (1986). Urban Hydrology for Small Watersheds, Technical Release 55. Washington,
DC.
Southwest Wetlands Interpretive Association (SWIA) (2001). Final Environmental Impact Statement and
Environmental Impact Report for the Goat Canyon Enhancement Project. Imperial Beach, CA.
Syvitski, J. P., Morehead, M. D., Bahr, D. B., & Mulder, T. (2000). Estimating fluvial sediment transport: the
rating parameters. Water Resources Research, 36(9), 2747-2760.
28
-------
Urbonas, B., and Stahre, P., (1993). Stormwater Best Management Practices and Detention. PTR Prentice Hall,
Englewood Cliffs, New Jersey.
29
-------
APPENDIX A. HYDROGRAPHS AND HYETOGRAPHS FOR ALL EVENTS
Table Al. Summary of storms and partitioning of rainfall into daily totals for analysis and modeling. The indicates events that were not included in
further analysis but were included for reallocation of rainfall. El, E2 or E3 indicate the events retained for analysis. Observed and revised rainfall are from
the Hormiguitas gage (RG.HM).

Daily rainfall
(mm)
Event total rainfall
(mm)
Event start
Event end
Maximum Intensity (mm)
SCS Storm
Type
Storm 1




15 min
1 hr
6 hr

2/27/2014*
1.3
1.3
2014-02-27 07:40
2014-02-28 00:00
0.75
0.75
1
-
El: 2/28/2014
19.7
12.2
2014-02-28 00:00
2014-02-28 15:50
2.75
5.75
9.75
12T2
E2: 3/1/2014
10.5
7.5
2014-02-28 15:50
2014-03-01 00:00
6.0
7.25
7.75
6T2
E3: 3/2/2014
0.5
7.5
2014-03-01 00:00
2014-03-01 15:57
3.25
6.0
7.25
-
3/3/2014*
0
3.5
2014-03-01 15:57
2014-03-02 12:13
1.5
1.5
3
-
Total
32.0
32.0















Storm 2








2015-02-28*
1.3
1.3
2015-02-28 11:26
2015-03-01 00:00
1.25
1.25
1.25
-
El: 2015-03-01
29.5
23.3
2015-03-01 00:00
2015-03-01 22:19
1.75
5.75
16.0
24T1
E2: 2015-03-02
5.2
9.2
2015-03-01 22:19
2015-03-02 11:29
2.75
6.25
7.75
24T1
2015-03-03*
0.3
2.5
2015-03-02 11:30
2015-03-03 02:00
2.5
2.5
4.0
-
Total
36.3
36.3















Storm 3








2015-05-14*
1.5
1.5
2015-05-14 14:31
2015-05-15 00:00
0.5
0.5
1.75
-
El: 2015-05-15
22.5
22.5
2015-05-15 00:00
2015-05-15 13:14
4.25
10.25
19
12T2
Total
24.0
24.0















Storm 4








El: 2015-09-15
29.5
30.8
2015-09-15 10:47
2015-09-16 05:52
3.5
11.25
21
24T1
2015-09-16*
1.3
0
2015-09-16 05:52
--



-
Total
30.8
30.8















Storm 5








2016-01-04*
14.3
15
2016-01-04 02:27
2016-01-05 09:18
3
3.5
7
-
El: 2016-01-05
23.0
22.3
2016-01-05 09:18
2016-01-05 18:33
4.75
8.5
20
12T2
2016-01-06*
5.5
5.5
2016-01-05 18:33
2016-01-06 20:08
2.5
2.5
4.25
-
2016-01-07*
6.5
6.5
2016-01-06 20:08
2016-01-07 23:56
0.5
2
3.5
-
2016-01-08*
1
1
2016-01-07 23:56
2016-01-08 04:28
1
1
1
-
Total
50.3
50.3















Storm 6








2016-03-05*
1.0
0
-
-




El: 2016-03-06
5.5
6.5
2016-03-05 20:44
2016-03-06 08:55
0.25
0.5
1.25
16T2
30
-------
E2: 2016-03-07
23.0
23.0
2016-03-06 9:00
2016-03-08 10:07
4.75
8.5
16.25
12T2
2016-03-09*
0.2
0.2
2016-03-09 5:12
2016-03-11 15:55
0.25
0.25
0.25
-
2016-03-11*
3.8
3.8
2016-03-11 15:55
2016-03-11 17:36
1
2.25
3.75
-
Total
33.5
33.5















Storm 7








2016-04-07*
8.8
8.8
2016-04-07 6:37
2016-04-07 14:46
1.5
2.5
8
-
2016-04-08*
1.2
1.2
2016-04-07 14:50
2016-04-08 7:18
0.25
0.75
1
-
2016-04-09/10*
3.8
3.8
2016-04-09 19:44
2016-04-10 4:47
0.75
1.25
3.25
-
Total
13.8
13.8















Storm 8








2017-01-17/18*
1.2
0
-
--




El: 2017-01-19
11.8
13.0
2017-01-17 3:49
2017-01-19 12:00
2.75
6
11.75
6T2
E2: 2017-01-20
28.0
28.0
2017-01-20 2:30
2017-01-21 23:00
4
6
14.25
16T1
2017-01-22*
11.0
11.0
2017-01-22 17:50
2017-01-22 23:26
1
3.5
11
-
2017-01-23*
13.0
13.0
2017-01-23 3:05
2017-01-24 06:00
5.75
7
11
-
Total
65.0
65.0















Storm 9








El: 2017-02-17
31.0
33.2
2017-02-17 17:50
2017-02-18 20:00
4.25
9.5
30.5
6T1
2017-02-18*
8.0
5.8
2017-02-18 20:00
2017-02-18 23:15
1.75
3
3
-
2017-02-19*
2.3
2.3
2017-02-19 2:59
2017-02-19 12:20
0.75
1
1.75
-
2017-02-22*
0.5
0.5
2017-02-22 1:18
2017-02-22 7:58
0.25
0.25
0.25
-
Total
41.8
41.8















Storm 10








2017-02-26*
2.0
0
-
--




El: 2017-02-27
74.5
83.0
2017-02-26 8:44
2017-02-28 13:28
1.75
6.25
33.25
24T1
2017-02-28*
6.5
0
-
--



-
Total
83.0
83.0






Table A2. Summary of storm events defined in Table 2.3. Source refers to which gage was used as the final observed data.

Event Date*

Rainfall
(mm)

Peak Discharge (cms)

Total Runoff (mm)

Runoff Ratio (Q/P)

Event

Source





PT
IBWC

PT
IBWC

PT
IBWC




Storm 1

2014-02-28

12.25

1.13
0.05

0.27
0.02

0.02
0.00

El

PT
31
-------

2014-03-01

7.50

0.50
1.54

0.13
0.33

0.02
0.04

E2

IBWC

2014-03-02

7.50

0.77
6.14

0.26
1.08

0.03
0.14

E3

IBWC
Storm 2

2015-03-01

23.25

3.36
-

1.36
-

0.06
-

El

PT

2015-03-02

9.25

1.43
-

0.48
-

0.05
-

E2

PT
Storm 3

2015-05-15

22.50

19.46
-

5.93
-

0.26
-

El

PT
Storm 4

2015-09-15

30.75

5.27
-

6.40
-

0.21
-

El

PT
Storm 5

2016-01-05

22.25

17.72
9.31

3.76
13.76

0.17
0.62

El

PT
Storm 6

2016-03-06

6.50

1.03
0.00

0.93
0.00

0.14
0.00

El

PT

2016-03-07

23.00

1.78
5.07

1.81
4.23

0.08
0.18

E2

IBWC
Storm 8

2017-01-19

13.00

-
5.37

-
2.57

-
0.20

El

IBWC

2017-01-20

29.25

-
6.86

-
18.66

-
0.64

E2

IBWC
Storm 9

2017-02-17

33.25

0.92
11.16

1.02
7.03

0.03
0.21

El

IBWC
Storm 10

2017-02-27

83.00

16.69
14.45

42.07
43.44

0.51
0.52

El

CAMERA
No PT data for storm 7, IBWC rating curve discharge was zero.
Table	A2	generated from:
https://github.com/kristaniguclii/EPA	Events Report TJ LLCW Scripts/blob/master/Table2.2 2.3 EventsReport generate.
R (generate	tables	from each events script),
https://github.com/kristaniguclii/EPA	Events Report TJ LLCW Scripts/blob/master/Table2.2 EventsReport html format.
R (format as html table)
Storm 2: 2015-03-01 to 2015-03-03
This storm had three distinct storm hydrographs (Figure Al). We separated them into two storms, one for
2015-03-01 and one for 2015-03-02 (Table 2.3). The third event was small and was excluded from the
model calibration and validation. The rainfall was closest to a 24-hour, Type I storm (Figure A2).
32
-------
B)
- TJE Naval
— PT
.£^V_

°) 1



1 	

A tx.
2015-03-01
2015-03-02
Date
2015-03-03
Figure A1. Storm 2. 2015-03-01 to 2015-03-03, with A) cumulative rainfall, B) pressure, including atmospheric pressure from
the weather stations (upper green line), adjusted atmospheric pressure (lower green line), and pressure from the PT (blue), C)
water stage, and D) discharge. The vertical dashed lines indicate where events were defined to start and end for purposes of
reallocating rainfall and runoff data (Table 2.3). El.PT and E2.PT indicate the two events that were retained for analysis.
https://github.com/kristaniguclii/EPA Events Report TJ LLCW Scripts/blob/master/figure 2.07 storm2 PT 2015 03 01
KT edits04202017.R
Storm type comparison March2015
10	15
time (hours)
20
Figure A2. Cumulative rainfall amount, normalized to event total rainfall, for the two events in March 2015 (Storm 2).
S tor in 3: 2015-05-15
Type 1
Type 2
Type 3
Event 1
Event 2
33
-------
This storm had one hydrograph event that occurred in the middle of the day, and the observed rainfall and
runoff time series were not changed for model input (Figure A3). This storm was an outlier for peak
discharge. The storm has higher maximum intensity than the Type II storm (Figure A4).
E
E
c
03
£Y.
E
ZJ
-------
Storm type comparison May2015
O
o
0	5	10	15	20
time (hours)
Figure A4. Cumulative rainfall amount, normalized to storm total rainfall, for the one event in May 15, 2015. This storm was
an outlier for peak discharge.
Storm 4: 2015-09-15
This storm has one hydrograph event (Figure A5). The event occurred on 2015-09-15 and was not
changed from the observed rainfall and runoff time series. A second event, on 2015-09-16, occurred after
rainfall stopped and is not shown. The reason for the second peak is not known but is likely due to
precipitation in the watershed not captured by the rain gages. Subsequent tests of the PT suggest that the
instrument deployed during this storm shows spontaneous fluctuation, and was replaced for subsequent
events. The rainfall was 24-hour type I (Figure A6).
Type 1
Type 2
Type 3
Event 1
35
-------
O -t
B) — TJE Naval
— PT
r*—^ _







o
CN
d
o
o
o
o
o
CO
CD
CM
o
2015-09-15
2015-09-15
Date
Figure A5. Storm event #4, 2015-09-15, with A) cumulative rainfall, B) pressure, including atmospheric pressure from the
weather station (upper green line), adjusted atmospheric pressure (lower green line), and pressure from the PT (blue), C) water
stage, and D) discharge. Vertical lines indicate the start and end of the one event retained for model validation. Prepared with
https://athub.com/Tmstaniguchi/EPA Events Report TJ LLCW Scripts/blob/master/figure 2.11 stonn4 PT 2015 09 15
KTedits04202017.R
Storm type comparison SEP2015
O
Type 1
Type 2
Type 3
Event 1
time (hours)
36
-------
Figure A6. Cumulative rainfall amount, normalized to storm total rainfall, for the one event in September, 2015. 24-hour type
I.
Storm 5: 2016-01-05
This storm has one hydrograph event on 2016-01-05 (Figure A7), so no reallocation of rainfall or runoff
data were performed. The rainfall most closely matched the 12-hour Type II storm (Figure A8).
zl.PT

TJE Naval
PT
T
-W"T>
0) "" IBWC Stage
* Visual Stage
- - - ¦




*





Dt ~~ PT
u' "" IBWC
* Visual Q calc

J
/

*
2016-01-04	2016-01-05	2016-01-06	2016-01-07	2016-01-08	2016-01-09
Date	Figure
A7. Storm event #5, 2016-01-05, with A) cumulative rainfall, B) pressure, including atmospheric pressure from the weather
station (upper green line), adjusted atmospheric pressure (lower green line), and pressure from the PT (blue), C) water stage,
and D) discharge. Vertical dashed lines indicate the start and end of the one event using IBWC BUBL stage. The vertical solid
lines indicate the start and end of one event using the PT and was retained for model validation. Prepared with
https://github.com/kristaniguclii/EPA Events Report TJ LLCW Scripts/blob/master/figure 2.13 storm5 PT 2016 01 04
KTedits04172017.R
37
-------
Storm type comparison JAN2016
Type 1
Type 2
Type 3
Event 1
o
0	5	10	15	20
time (hours)
Figure A8. Cumulative rainfall for January, 2016. This stonn was an outlier for peak discharge.
Storm 6: 2016-03-06
This storm has one hydrograph event on 2016-03-06 and one hydrograph event on 2016-03-07 to 2016-
03-08 (Figure A9). The PT gave erratic measurements during the second event that did not correspond
closely with rainfall, so the IBWC BBLR and ICBW-PT rating curve were used for that event. The rainfall
did not match any storm type, but the peak intensity corresponded with a 16-hour, Type II storm (Figure
A10).
38
-------
E
E
L
"ro
a:
E
Zi
O
2
CL
 — TJE Naval


— PT








C) -- IBWC
* Visual
— PT

in -

/

E
O

GO
Q
o -


1 '
i r
t i'
2016-03-06
2016-03-07
2016-03-08
2016-03-09
Date
Figure A9. Storm event #6, 2016-03-06, with A) cumulative rainfall, B) pressure, including atmospheric pressure from the
weather stations (upper green line), adjusted atmospheric pressure (lower green line), and pressure from the PT (blue), C) water
stage, and D) discharge. Vertical lines indicate the start and end of the one event retained for model validation. The PT data for
E2 were not used due to erratic measurements that do not correspond to the rainfall, so the IBWC BBLR data and IBWC-PT
rating	curve	was	used	instead.	Prepared	with
https://github.com/kristaniguclTi/EPA Events Report TJ LLCW Scriots/blob/master/figure 2.15 stonn6 PT 2016 03b K
Tedits04172017.R.
Storm Type comparison March2016
£=
13
O
c
CD

to
=5
CO
o
iO
o
o
CN
o
o
o
Type 1
Type 2
Type 3
Event 1
10	15
time (hours)
39
-------
Figure A10. Cumulative rainfall for March, 2016.
Storm 7: 2016-04-09
This storm did not have recorded runoff at the PT, despite having significant rainfall (Figure All). A
malfunction of the PT must have occurred during this storm. Additionally, IBWC rating curve gave values
of zero for this storm due to low stage measurements recorded from the bubbler. This storm was not
included in subsequent analysis.
DC
Z!
o
TJE Naval
2016-04-09
2016-04-11
Figure All. Storm event #7, 2016-04-09, with A) cumulative rainfall and B) pressure, including atmospheric pressure from
the weather station (upper green line), adjusted atmospheric pressure (lower green line), and pressure from the PT (blue). No
apparent discharge event captured with the PT, IBWC rating curve had discharge values of zero. Figure generated with
https://github.com/kristaniguclii/EPA Events Report TJ LLCW Scripts/blob/master/figure 2.17 stonn7 PT 2016 04 KT
edits04172017.R
Storm 8: 2017-01-19
A malfunction of the PT occurred during this storm. According to the IBWC rating curve, this storm had
three major storm hydrograph (Figure A12). We retained two storms for analysis, one for 2017-01-19
and one for 2017-01-20 (Table 2.3). The third storm was erratic and didn't correspond well with rainfall
and was excluded from the model calibration and validation. Storm was type 2, 6 hour (Figure A13).
40
-------
E
c
CD
Q£
E
ZS
o
A> E1 IBWC
E2.IBWC
r1
s
y

/

E^
-------
Storm type comparison January 2017
a
=s
o
TO
'c
"to
1	
CD
_>
TO
Z5
E
Z5
O
CO
o
CO
o
o
CNj
o
p
o

¦ Type 1

- Type 2

- Type 3

- Event 1
¦
Event 2
15
20
time (hours)
Figure A13. Cumulative rainfall for January 2017.
Storm 9: 2017-02-17
This storm had one distinct stonn hydrograph (Figure A14). The PT housing was damaged during this storm and gave erratic
measurements. Discharge calculated from the IBWC rating curve was used for the model calibration and validation. The
IBWC peak discharge (-10 m3 s"1) matched well with the observed discharge (-15 m3 s"1). Storm was 6 hour type I (Figure
A15).
42
-------
o
Barologger
PT
CD
O
CN
O
03
CD
A
IBWC
PT
Visual

o


— IBWC y.


D) — PT


° Visual i


1 \


f ^

_
. A ... ° '
!
2017-02-18	2017-02-19
Date
Figure A14. Stomi event #9, 2017-02-17, with A) cumulative rainfall, B) pressure, including atmospheric pressure from the
barologger (lower black line), adjusted atmospheric pressure (upper black line), and pressure from the PT (blue), C) water stage
from the PT (solid black line) and IBWC bubbler (dashed black line), and D) discharge from the PT and IBWC rating curve.
El.IBWC indicates the one event was retained for the model and validation using the IBWC rating curve. Figure generated
with
https://github.com/kristaniguclii/EPA Events Report TJ LLCW Scripts/blob/master/figure 2.19 stonn9 IBWC visual 20
17 02 KTedits05012017.R
43
-------
Storm type comparison Feb 17 2017
a
=s
o
TO
'c
"to
1	
CD
_>
TO
Z5
E
Z5
O
CO
o
CO
o
o
CNj
o
p
o
Type 2
Type 3
Event 1
time (hours)
Figure A15. Cumulative rainfall for February 17, 2017.
Storm 10: 2017-02-27
This storm was the largest recorded observed storm and had one distinct storm hydrograph (Figure A16). There was no data
from the PT, but a field camera was placed at the PT location and recorded stage every 15 minutes. The IBWC rating curve
was developed from this event. Discharge calculated from the field camera was used for the model calibration and validation.
Storm was 24h type I (A17).
44
-------
IBWC
Camera
NoPT
IBWC
Camera
NoPT
Obs Vel - Q
Figure A16. Storm event #10, 2017-02-27, with A) cumulative rainfall, B) stage recorded by the IBWC bubbler (dashed black
line) and stage recorded by the field camera (solid black line), and C) discharge from the field camera and IBWC rating curve.
IBWC rating curve was based on this event. Discharge from the field camera matched closely with observed discharge and was
used	in	model	calibration	and	validation.	Figure	generated	with
https://github.com/kristaniguclii/EPA Events Report TJ LLCW Scripts/blob/master/figure 2.20 stonnlO IBWC visual 2
017 0227 KTedits05012017.R
Figure A17. Cumulative rainfall for February 17, 2017.
45
-------
APPENDIX B. HYDROGRAPHS DURING SSC MEASUREMENTS
to
E
0

o
o
CD
CM
CD
CN
CD
CN
CD
CM
CD
CM
co
CM
CD
CM
CD
CM
CD
CM
CD
CM
CD
CM
CD
CN
O
o
o
o
o
o
o
o
o
o
o
CM
O
CM
O
CM
O
CM
O
CM
O
CM
O
CM
o
CM
O
CM
O
CM
O
CM
O
CM
O
co
o
co
o
co
o
co
o
co
o
co
o
co
o
co
o
co
o
co
o
co
o
Figure Bl. IHfydrograph and SSC samples for 02/28/2014 - 03/01/2014. Figure generated from
https://github.com/kristaniguchi/EPA Events Report TJ LLCW Scripts/blob/master/Figure 3.2 EventsReport SSC 03012
014.R
E
o
-------
https://github.com/kristaniguclii/EPA Events Report TJ LLCW Scripts/blob/master/Figure 3.3 EventsReport SSC 03012
015.R
o
CM
£
O
-------
to
£
O
-------
49
-------
APPENDIX C. SEDIMENT TRAP TEXTURE ANALYSES AND TRAP
EFFICIENCY
Table CI. Particle size data summary for the Goat Canyon sediment traps. All sample depths were 0-3 ft and mean grain size description was
"fine sand" for all samples. Samples SSI through GC8 are from AMEC (2007). Samples from "Avulsion Basin #1 to Canyon Basin #2 are from
de Temple et al. (1999).
Sample
location
Sample
ID

Particle size distribution, percent




Median
grain size
mm
Gravel
Sand
Silt
Clay
Silt + Clay
Total Sand
Coarse
Med.
Fine


Sorted Pile
SSI
0.098
0
0
17.11
45.79
31.31
5.78
37.10
62.9
SS2
0.101
0
0
20.07
44.32
30.91
4.71
35.61
64.39
SS3
0.094
0
0
12.39
50.82
33.62
3.17
36.79
63.21

Mean ±
sd
0.098±
0.003
0
0
16.5±3.9
47±3.4
31.9±1.5
4.6±1.3
36.5±0.8
63.5±0.8











Native Pile
NS1
0.122
0
0
27.55
43.05
26.16
3.24
29.4
70.60

NS2
0.146
0
0
30.01
43.93
22.81
3.25
26.05
73.95

NS3
0.152
0
0
28.3
47.04
20.66
4
24.65
75.34

Mean ±
sd
140± 0.016
0
0
28.6±1.3
44.7± 2.1
23.2± 2.8
3.5± 0.4
26.7± 2.4
73.3± 2.4











Upper
Catchbasin
GC1
0.090
0
0
7.60
53.97
33.69
4.74
38.43
61.57

GC2
0.075
0
0
3.48
47.67
43.19
5.65
48.85
51.15

GC3
0.085
0
0
3.95
54.83
36.4
4.82
41.22
58.78

GC4
0.075
0
0
10.08
40.33
42.62
6.96
49.58
50.42

GC5
0.069
0
0
5.39
41.06
47.57
5.98
53.54
46.45

Mean ±
sd
0.079±
0.008
0
0
6.1±2.7
47.6±6.9
40.7±5.6
5.6±0.9
46.3±6.3
53.7±6.3
Lower
catchbasin
GC6
0.082
0
0
9.04
45.02
40.47
5.47
45.94
54.06

GC7
0.094
0
0
3.38
61.31
30.79
4.53
35.32
64.68

GC8
0.102
0
0
18.73
43.16
33.09
5.02
38.11
61.89

Mean ±
sd
0.092±
0.010
0
0
10.4 ±7.8
49.8 ± 10
34.8 ± 5.1
5±0.5
39.8±5.5
60.2±5.5











Avulsion basin
#1
-
0
-
-
-
-
-
37.0
63.0
Avulsion basin
#2
-
0
-
-
-
-
-
13.1
89.6
50
-------
Silt basin
-
-
0
-
-
-
-
-
39.8
60.2
Canyon basin
#1
-
5.8*
-
-
-
-
-
16.1
78.1
Canyon basin
#2
-
5.9*
-
-
-
-
-
3.4
90.7
*Coarser fractions underestimated due to sampling methods.
The settling velocity (co) for each sediment size was estimated using the equations in the Reservoir
Sedimentation Handbook referring to the Rubey (1933) equation:
[I636(pf ~ p)^ + Vf5—3^
500rf	(Ci)
where co = terminal fall velocity (m s"1); ps = sediment density (kg m"3); p = density of water (kg m"3),
assumed to be 1000 kg m"3; p = dynamic viscosity of water (N*s m"2), assumed to be 1.31xl0"3 N*s/m~2,
and d = particle diameter (m).
The critical settling velocity (coc) of the sedimentation basin was calculated as:
coc = Q/A
(C2)
where coc = critical settling velocity (m/s), which is the velocity of the slowest particle of the basin that
will be 100% removed (Morris and Fan, 1998); Q = design discharge or inflow
(m3 s"1), and A = surface area of the sediment basin (m2).
51
-------
Table C2. Sediment removed from traps (Tons Removed), annual trap efficiency, and corrected sediment load from the watershed by size
class.

Removal Date
Tons Removed
Eann n=l
Eann n=3
Corrected Load (tons)
2006

Medium sand (a)
1947
1.00
1.00
1947

Fine sand (b)
15194
1.00
1.00
15194

Silt (c)
12991
0.99
1.00
12992

Clay (d)
1788
0.36
0.40
4508

Total
31920


34642

Total without Clay
30132


30133
2007

Medium sand (a)
1947
1.00
1.00
1947

Fine sand (b)
15194
1.00
1.00
15194

Silt (c)
12991
0.99
1.00
12992

Clay (d)
1788
0.53
0.61
2946

Total
31920


33079

Total without Clay
30132


30133
2008

Medium sand (a)
3115
1.00
1.00
3115

Fine sand (b)
24310
1.00
1.00
24310

Silt (c)
20786
0.96
1.00
20815

Clay (d)
2860
0.16
0.18
16339

Total
51072


64580

Total without Clay
48212


48241
2009

Medium sand (a)
3536
1.00
1.00
3536

Fine sand (b)
27592
1.00
1.00
27592

Silt (c)
23593
0.95
1.00
23673

Clay (d)
3246
0.22
0.23
14148

Total
57967


68949

Total without Clay
54721


54801
2010

Medium sand (a)
4284
1.00
1.00
4284

Fine sand (b)
33427
1.00
1.00
33427

Silt (c)
28581
0.97
1.00
28609

Clay (d)
3933
0.29
0.31
12615

Total
70224


78935

Total without Clay
66291


66320
2011

Medium sand (a)
3951
1.00
1.00
3951

Fine sand (b)
30833
1.00
1.00
30833

Silt (c)
26364
0.93
0.99
26764
52
-------

Clay (d)
3627
0.34
0.39
9416

Total
64776


70965

Total without Clay
61149


61549
2012

Medium sand (a)
3505
1.00
1.00
3505

Fine sand (b)
27349
1.00
1.00
27349

Silt (c)
23385
0.99
1.00
23388

Clay (d)
3218
0.67
0.75
4271

Total
57456


58513

Total without Clay
54238


54242
Source: https://github.com/kristaniguchi/EPA Events Report TJ LLCW Scripts/blob/master/Table4.2 trap efBciencv.R
a.	6.1% Medium sand: 0.25 - 0.5 mm (mean = 0.375 mm)
b.	47.6% Fine sand: 0.125 - 0.250 mm (mean = 0.1875 mm)
c.	40.7% Silt: 0.0039 - 0.0625 mm (mean = 0.0332 mm)
d.	5.6% Clay: 0.00098 - 0.0039 mm or <3.9 um (mean = 0.00244 mm)
The mean grain size diameter was used to calculate the trap efficiency for each size class.
53
-------
Original data figures and reports from AM EC (2007)
NATIVE
STOCKPILE
ciwrt Cnnyoo sampling locelKxis
IJppw and Lower Retention 8as in and Stockpile Areas
Figure CI. Sites analyzed for particle size by AMEC (2007).
» Avulsion Basin SI
••^Avulsi:
QpK	1 , frrr-	0
Canyon Basin #
54
-------
Figure C2. Sites analyzed for particle size by DeTemple et al. (1999). Sites are not located in the current
sediment traps, but were taken in the Tijuana Estuary prior to constructing the sediment traps.
Table C3. Mean soil particle size data from AMEC (2007).
Table 2. Summary of Mean Soil Particle Size Data
Soil
Sampling
Area
Number
of
Samples
Percent
Sand1
(Mean ± SD)
Percent
Fines2
(Mean ± SD)
Median
Grain Size
(mm)
(Mean ± SD)
Upper Basin
5
53.7 ± 6.3
46.3 ± 6.3
0.079 ±0.006
Lower Basin
3
60.2 ± 5.5
39.8 ± 5.5
0.092 ±0.010
Native Stockpile
3
73.3 ± 2.4
26.7 ±2.4
0.140 ±0.016
Sorted Stockpile
3
63.5 ± 0.8
36.5 ± 0.8
0.098 ± 0.003
1 - particle size range from 0.063 to 3.363 mm.
* - particle size below 0.063 mm.
mm - mlimeters
SD - slandafd deviation
Table C4. Raw data of particle size in the Goat canyon sediment traps, used to calculate the means in
Table CI (from Table C-l in AMEC 2007).
55
-------




IM1»
Pvtici* Stii OiiPlbirttoft. «i ptfcint
5611
Mma
CoBacficm
S^npJhi D
Daptih, «_

Gr*ifi Sua

SMd S4t*


ft
PtrcMt
LocMkm



M
Or Ml
Com ] Hitfun [ nil*
»¦
a*

Sirtd

SSI
0-3
F)m until
0.098
0 00
0.00
ir.ii
45.79
kit
5.71
37.10
62.90

SS2
0-3
Frit ynd
Q.KM
0.00
CL00
20.07
44.32
30-91
4.71
35,61
64-39
SmdPte
S£3
CW
F nfi
O.OM
0.00
0.00
12.39
AG
33.62
3.17
38.79
6321

Uw


O.OM
oo
0 0
165
47.0
31.9
a
365
63.50

Standard Dertjfton

0 003
00
00
3.9
14
iS
1.3
OB
078
i Sort«t3 Hi*

*•*



*5


ym


MSI
0-3
Ptm itno
0.122
0.00
0,00
27 55
4 3,05
26,»6
3.24
29.40
70.60

NS2
0-3
F« band
0.1*6
0.00
0,00
30 01
4193
22 JM
3.25
26.05
73.96
(UrwMd)
NS3
«V3
Frt uu
Q.152
0.00
0.00
28.30
47.04
20.66
4.00
2 4..66
75.34



0,140
00
00
21.6
44,7
232
15
2*7
7X30

Standard DeriJHm

0,016
00
00
1.3
Zi
2£
OA
2,4
2,43

QC1
0-3
Firm urti
0.090
0.00
0.00
7.60
53.97
33-69
4,74
38.43
61.57

OC2
CV3
F« until
0.076
0.00
0.00
3.48
47.87
4119
5.66
48J5
51.15

GC3
0.3
yr>d
0.06*.
0.00
o.oo
3 95
54 83
36.40
4J2
41,22
58v78

OC4
0-3
F«* ftM
0.075
0.00
0.00
10.08
40.33
42-62
6.96
49.58
50.42

OCS
0-3
F« fc*K3
0.069
000
0:00
539
41.06
47.57
5 98
5154
46.46

Mhti


0.079
0.0
0 0
1.1
47.6
40.7
il
46.3
5167

SUKtardDi
riJOOTi

DQG-9
ao
QJQ
2J
&9
5l6
0.9
6.3
t27
Siatttttc jI $iQ«iAcM«c* EUppw rt Lom
ir CtfcMuiifli?
MS


MS
*3


MS


QC6
0-3
F« wti
0.062
0.00
0.00
9.04
45.02
40,47
5^7
45-94
54 06

OCT
0-3
F« urti
0.094
ox
0.00
3.38
61.31
30.79
4.53
35-32
64.68
CMCfCM
oca
0^3
Fru urtJ
0.KE
0.00
0.00
18.73
43.16
33J»
5.02
38.11
61J»
u»an


0.092
00
OO
10.4
49,8
Ml
50
39J
60.21

Swdvd OtMlafl

0.010
oo
00
7.8
100
11
0.5
5.5
5.51
CJleffciwi












Sotc<»s
u*an


0.0S*
0 000
0000
7.706
48 420
384 78
5.396
41874
49*9

Standard D»»4®or>

0.011
0000
0000
5.141
7533
1872
0 803
6.528
6.53
Figure C3. Photograph of sedimentation basin installed in winter 2013-14, Photo was taken on January 15, 2016.
56
-------
Figure C4. Sediment sample taken at the retention basin 011 January 15, 2016.
57
-------
APPENDIX D. LINKS TO DATA AND SCRIPTS
Figure 1.1 source code:
https://github.com/kristaniguchi/EPA Events Report TJ LLCW ArcMap/blob/master/Figure 1.1 eve
nts report wtshd map.mpk
Figure 2.1 Source code
https://github.com/kristaniguchi/EPA Events Report TJ LLCW Scripts/blob/master/figure 1.2 hypso
metric dem lOObins.R
Elevation bins generated using:
https://github.com/kristaniguchi/EPA Events Report TJ LLCW ArcMap/blob/master/Figure 1.2 hyp
osometric curve dem clip wtshd forRscript.mpk
Figure 2.2 Source code:
https://github.com/tbiggsgithub/EPA Events Report TJ LLCW Scripts/blob/master/precip data QC.
R
Figure 2.3 Source code:
https://github.com/tbiggsgithub/EPA Events Report TJ LLCW Scripts/blob/master/precip data QC.
R
Figure 2.4 Source code:
https://github.com/tbiggsgithub/EPA Events Report TJ LLCW Scripts/blob/master/precip sum over
storm events.R
Table 2.3 Source code:
https://github.com/kristaniguchi/EPA Events Report TJ LLCW Scripts/blob/master/Table2.2 2.3 Ev
entsReport generate.R (generate tables from each events script),
https://github.com/kristaniguchi/EPA Events Report TJ LLCW Scripts/blob/master/Table2.2 Events
Report html format.R (format as html table)
Figure 2.6 Source code:
https://github.com/kristaniguchi/EPA Events Report TJ LLCW Scripts/blob/master/figure 2.02 stor
ml PT 2014 03 01 KTedits04172017.R
Figure 2.7 Source code:
https://github.com/kristaniguchi/EPA Events Report TJ LLCW Scripts/blob/master/figure 2.04 ibwc
ratingcurve.R
58
-------
Figure 2.9 Source code:
https://github.com/tbiggsgithub/EPA Events Report TJ LLCW Scripts/blob/master/figure 2.5 rainfal
1 runoff SCS CNR
Figure 2.10 Source code:
https://github.com/tbiggsgithub/EPA Events Report TJ LLCW Scripts/blob/master/figure 2.6 Pmax
vs Qmax.R
Table 2.4 Source code:
https://github.com/kristaniguchi/EPA Events Report TJ LLCW Scripts/blob/master/Table3.1 3.3 Ev
entsReport generate.R and
https://github.com/kristaniguchi/EPA Events Report TJ LLCW Scripts/blob/master/Table3.1 Events
Report html format.R
Table 2.5 source code:
https://github.com/tbiggsgithub/EPA Events Report TJ LLCW Scripts/blob/master/Table3.2 EventsR
eport html format.R
Figure 2.11 Source code:
https://github.com/tbiggsgithub/EPA Events Report TJ LLCW Scripts/blob/master/regression model
s SSC vs OR
Figure 2.12 Source code:
https://github.com/tbiggsgithub/EPA Events Report TJ LLCW Scripts/blob/master/figure 4.1b break
dates.R
59
-------