Jordan River Project Water Use Plan Lower Jordan River Inflow Monitoring
Reference: JORMON#1 Lower Jordan River Inflow Monitoring:Final Report for 2005-06
Study Period: December 2005 – October 2006
Robert Hudson, Ph.D., P.Geo., HYDROID Geoscience Ltd.
October 2006
1
Lower Jordan River Inflow Monitoring: Final Report for 2005-06
Submitted to Ian Dodd and Alf Leake, BC Hydro.
By Robert Hudson, Ph.D., P.Geo., HYDROID Geoscience Ltd.
2 Executive summary On behalf of BC Hydro, HYDROID Geoscience Ltd. undertook a study of local inflows into the Jordan River below Elliot dam as part of a study to evaluate the effectiveness of a steady release of 0.25 m3/s at improving fish habitat. The study that we proposed and carried out went beyond the original terms of reference in that we measured the flow not only in the mainstem channel but also in the tributary channels in order to determine whether or not the mainstem channel is losing water through leakage or gaining water by groundwater inflow. The following summarizes our methods and findings. • Hydrometric data were collected using salt dilution streamflow gauging methods, which are well suited to measuring flow in steep turbulent and irregular channels where current metering fails. While not in the RISC standards, these methods are well known to the scientific community. • Rating curves were developed using an innovative method that is fully documented in this report. • Measurement of streamflow in the mainstem channel was extremely challenging because the flow is under-fit due to the removal of flow by the upstream dams. Conditions violated the assumptions of both current metering and salt dilution most of the time. • In spite of the above difficulties we achieved 5% measurement precision and 5 – 10% rating accuracy. We consider this exceptional given the conditions. • A flow budget was constructed by adding up total tributary flows and comparing the totals to measured flow between points on the mainstem channel. Some of the tributary flows were estimated based on the measured flows. • The flow budget indicated that the mainstem channel is not only watertight, but low flows are sustained by groundwater inflow in the order of 0.01 m3/s. • The flow budget and rating curves need to be validated early in the 2006-07 contractyear by installing an additional stream gauge and collecting additional flow measurements at all sites. These elements are required to minimize uncertainties in the flow budget. • Notwithstanding the above point, we believe that the 0.25 m3/s flow release will be entirely measurable throughout the length of the mainstem below Elliot dam and can be implemented in the summer of 2007. • The primary objectives for the project can be partially supported with existing data, but further monitoring with additional installations is require before the management questions can be answered with certainty.
3 Introduction: This is a report of 2005-06 contract-year findings of the study of the flow characteristics and local tributary inflows of the lower Jordan River. The Jordan River, located on the southwest tip of Vancouver Island is the subject of a fisheries rehabilitation project. BC Hydro operates a series of dams to generate electricity. At the lowest dam (Elliot) the water is diverted into a penstock such that effectively, there is no flow in the main channel below the dam. Recently, the Jordan River Water Use Plan was approved and implemented, which included the installation of a low-flow port from Elliott Dam to deliver 0.25 cms to the Lower Jordan River, and a monitoring program to assess its effectiveness on fish production. As part of the monitoring program we undertook a study of low flow characteristics of the lower Jordan and its tributaries. In this case, low flow is defined as the “normal” flow regime in the absence of spillway flows, most or all which is derived from the tributaries that empty into the Jordan River below Elliot dam. Study Objectives as stated in the JOR Terms of Reference: The primary management questions regarding the natural inflows below Elliott Dam are: 1. How accurate were the assumptions of local inflows used for WUP recommendations? 2. What implications, if any, are there on the WUP recommendations based on revised inflow data? 3. What are the reasons for the differences, if any, between the monitored and assumed inflows? The main problem with the estimated inflows is that they are based entirely on modeled data. There are no real measured time series data for inflows either above or below Elliot Lake. Further, it is not known whether or not there is significant leakage out of the main channel, which could potentially negate any assumed benefits from the proposed flow release. These issues combine to create a high level of uncertainty surrounding the hydrology of the lower Jordan River. The objective of this monitoring study is to assess the performance of the 0.25 m3/s flow release to achieve fish habitat goals as per the WUP decision. This can be summarized as follows: 1. Using measured time series data, determine the accuracy of the modeled inflow data used to calculate the required flow release. 2. Determine whether or not there is significant flow leakage from the mainstem channel bed that could negate potential benefits from the flow release.
4 Jordan River: The Jordan River is an extremely un-natural channel (Figure 1). It is a large channel formed under an extreme precipitation regime with almost no flow in it most of the time. The Elliot dam effectively takes all the water derived from the upper part of the watershed and uses it for generation of electricity. There is no flow in the channel immediately below the dam except when the reservoir reaches capacity, after which the spillway begins to deliver water, or when BC Hydro opens the gates to lower the reservoir level. When either of these events occur, the flow is abruptly increased, some times to high flow conditions. As a result, there is very little gravel in the channel. Gravel that is recruited from tributary streams is transported away whenever high flow conditions occur. This leaves a channel bed that is composed almost entirely of large boulders (Figure 1). In the example shown (near the upper gauge on the mainstem channel), the wetted part of the channel consists of large pools that are connected together by flow that trickles from one pool to the next. These flows are not measurable by any standard method but can be estimated using first principles.
Figure 1: Upper main channel, JOR
5
Working map of lower Jordan River. The black lines are the 200, 400 and 600 metre contours. The solid red line outlines the tributary stream catchments or other collective drainage areas (e.g., M2 which incorporates several tributaries) that were monitored. The red dashed lines outline significant tributaries that were not monitored. The lower area that is outlined with a blue line is the residual area M1resid. The lower gauge, referred to as M1 includes the whole 18 km2 drainage.
6 Table 1: Area-elevation information, lower Jordan and its tributaries. (areas in km2)
Watershed Unit T1 T2 M2 T1.5 T2.5 M1residual M1 Status gauged gauged gauged ungauged ungauged ungauged gauged <200 elevation range (m., asl) 200-400 400-600 0.208 1.984 0.170 0.605 2.096 3.200 0.576 1.008 0.176 0.560 1.934 0.208 4.984 7.005 >600 1.200 0.540 2.928 0.288 0.840 4.956 Total 3.392 1.315 8.224 1.872 1.576 3.149 17.952
1.007 1.007
The watershed that supplies flow to the lower Jordan River is about 18 km2 in area (see map, next page). The elevation range is from near sea level to 740 metres asl. Six major tributary creeks were identified within that area. Three tributaries were selected for monitoring as well as three mainstem points. Watershed areas were determined by planimeter on a 1:40,000 scale TRIM base map, and were stratified by elevation and by drainage unit (Table 1). Methods: As stated in the BC Hydro terms of reference (appendix A) we established three stream gauges on the mainstem channel between Elliot dam and the power house. In addition to these we also established gauges on three tributary creeks. Each gauge consisted of the following equipment: Unidata data logger (either red or blue Starloggers) plus peripherals Water depth sensors: o Transducers: Unidata (JOR-M2) or Keller (JOR-T1 aka Sinn Fein) o Capacitive water depth probe (Unidata) (JOR-M1, M3, T2) For the sites using the capacitive probes a staff gauge was first established and the probe was attached to the support structure. For the transducer sites an innovative armoring system was used as described below. Armoring: for very active steep small streams where channel morphology changes frequently due to frequent high flow events, it can be very difficult to maintain a stream gauge. Not only do rating curves change but the in-stream instrumentation takes a beating. Sinn Fein Creek is one such stream. We used our standard approach to establish monitoring in this channel. We encased the transducer cable in wire rope and constructed an anchor out of a used grader blade (Figures 2 – 4). A post is cemented into the stream bank to act as both an anchor point for the transducer and a benchmark reference point for gauge leveling (Figure 5). The cable is vented at this point, usually using a junction box to house electrical connections and the desiccant tubes that are attached to the vent. Water levels were established using the standard method described in the RISC standards (MOE, 1998). At the sites equipped with staff gauges the gauges were
7 periodically checked against a fixed benchmark, and the level on the capacitive probe was routinely checked against the staff gauge reading. At the transducer sites with no staff gauge the position of the gauge was checked each time with the survey level and stadia rod by comparing the level of the fluorescent bolt on the Heavy Metal anchor (Figure 3) with the level of the benchmark on the anchor post (Figure 5). The water level was also read off the stadia rod using the fluorescent bolt as a reference point (Figure 4). Streamflow was measured using salt dilution gauging (mass balance method) as described by Hudson and Fraser (2006). The steep, turbulent tributary sites were particularly well suited to dilution gauging, but we also found this technique to be applicable even in the mainstem sites under low flow conditions. Conversely, there were no channel sections anywhere near where we wanted to set up the stream gauges where current metering could be applied without violating the assumptions of that method. Rating curves were developed using an innovative method based on the ChapmanRichards asymptotic-exponential curve (Sit, 1994) currently under development (Hudson, 2006). A rating curve is a stage – discharge relationship; that is, the stage or gauge height is plotted on the Y-axis against discharge on the X-axis. This is the standard approach because discharge is the independent variable and gauge height/stage/water level is the dependent variable. After having analyzed many gauge records we have found that an equation of the following form provides an excellent fit to most gauge data:
− b(Q ) ⎞ GH − Offset = a⎛1 − e ⎜ ⎟ ⎝ ⎠
c
(1)
where GH = gauge height, Offset = the Y-intercept of the curve, or the zero flow point, a = the asymptote, b and c are curve fitting parameters, and Q = stream discharge. The curve passes through the origin, however the benchmarks at a gauging site are established such that there is a positive offset between datum and the “zero flow” point. Therefore the offset in the equation is equal to the zero flow gauge height. We have developed a program to fit this curve to hydrometric gauge data. The parameters a, b and c are the curve fitting parameters; the “a” parameter is a horizontal asymptote, and the “b” and “c” parameters govern the shape of the curve. The user specifies the range for each parameter as well as the offset and the program optimizes the b and c parameters for the range of asymptotes and offsets specified. It then prints out a table containing the asymptotes, offsets and associated b and c parameters and the SSE. The offset (hence the zero flow GH) must be less than the minimum gauge height in a record and the sum of the offset and asymptote equal or greater than the maximum gauge height in the record. Once these criteria are met the best fit rating curve is the one with the lowest SSE. If future records indicate a large range between maximum and minimum gauge heights, the rating curve can be adjusted from the same table as long as the initial range of values specified by the user was large enough.
8
Figure 2: Heavy Metal stream anchor for water depth transducer.
Figure 3: Heavy Metal anchor ready for installation at Sinn Fein.
9
Figure 4: Stream anchor settles into the channel bed and protects the transducer from being crushed by boulders.
Figure 5: A post is cemented into the stream bank to act as a fixed anchor point as well as a benchmark reference for gauge leveling.
10
The inverted form, to be used for calculating discharge from a stage record is as follows:
⎛ ⎛ GH − offset ⎞1 / c ⎞ ⎟ ln⎜1 − ⎜ ⎜ ⎝ a⎟ ⎟ ⎠ ⎠ Q= ⎝ −b
(2)
The above equation is never extrapolated beyond the range of the data on which it was based. The high end of the rating curve is based on a linear relationship developed from the highest measured flows. This linear extrapolation is tangential to the rating curve at the highest measured discharge. The form of the linear equation is:
GH =
Q − B0 B1
(3)
where B1 is the slope of a line that is tangential to equation 1 and B0 is the intercept of the tangential line. The computational form of equation 3 can be written as follows:
Q = B0 + B1 ⋅ GH
(4)
The RISC standards include a graphical method to determine the “zero flow gauge height”. We tried this but found that it didn’t work. In the documentation of the method it is acknowledged that “the water level can and does fall below the actual zero flow level”. In our experience, the zero flow level estimated by this method is not the true zero point; the flow at this point is not zero, but un-measurable. We found that the true zero point was revealed over time when the streams receded to a flat-line state after an extended dry period. These conditions occurred at T1 and T2 in late August 2006. Using the minimum gauge height of the record as the offset, the rating curves were adjusted so as to minimize SSE for that offset. Results: The time series data are all in the attached spreadsheet. There is a continuous data record of 293 days starting on 12/15/05 until 10/03/06 for all sites except T3 where the instrument failed and did not yield any data. The site would have complimented the existing data set, as it was the uppermost tributary to supply flow to the mainstem channel.
11 Manual flow measurements: The manual measurements of discharge and the associated gauge heights (hydrometric data) collected under the current contract are summarized (Table 2). Salt dilution measurements were collected either in duplicate or quadruplicate for low flows; these are our standard replications depending on whether the data collection is routine or for the purpose of site calibration of mixing length. The data given in Table 2 represent the averages of the values obtained during each measurement. The measurement precision is 5% for low flows less than 0.04 m3/s and better for higher flows. This was determined as the difference between measurements over the mean value. Sources of error that contribute to this include instrument error, variability of conditions during measurement, errors in salt delivery etc. Note that the upper mainstem site (M3) was deliberately selected to represent the uppermost pool in the mainstem channel. There was no water in the channel above that point when we established the site. We could detect a trickle of flow from that pool to lower pools in the channel but could not actually measure the flow by any standard method. Flow was estimated by estimating the flow velocity and cross-sectional area of the trickle.
Table 2: Hydrometric measurements collected Dec 05 – Oct 06: all elevations in metres.
Site M1 M1 M1 M2 M2 M2 M3 T1 T1 T1 T1 T2 T2 T2 T2 Date/time 3/7/2006 11:52 6/1/2006 12:03 10/5/2006 9:15 3/7/2006 13:30 3/23/2006 15:53 5/31/2006 14:02 5/31/2006 12:00 12/14/2005 9:38 1/23/2006 10:15 3/23/2006 11:23 6/1/2006 10:02 12/13/2005 13:30 1/24/2006 12:17 3/23/2006 14:02 5/31/2006 15:26 Stage 0.683 0.125 0.026 0.562 0.463 0.191 0.397 0.666 0.571 0.384 0.293 0.426 0.399 0.239 Q (m3/s) 2.132 0.081 0.017 0.553 0.383 0.036 0.0003 0.040 0.886 0.465 0.015 0.013 0.123 0.089 0.005 Gauge Height 0.874 0.316 0.217 0.686 0.587 0.315 0.603 0.872 0.797 0.57 0.506 0.639 0.612 0.452 Staff Gauge 0.874 0.316 0.217 N/A S.G. N/A 2.400 2.400 2.400 2.400 0.324 0.198 0.224 0.158 N/A: GH = SG+0.5 2.121 1.726 1.827 1.988 0.206 0.206 0.226 0.186 0.213 0.213 0.213 0.213 BM elevation Level to BM N/A: GH = SG 1.185 1.185 1.185 0.512 0.323 0.456 1.011 0.921 1.326 Level to Water Stage – GH offset 0.191 0.191 0.191 0.124 0.124 0.124
0.006 0.139 0.112 -0.048
Measured discharge data collected by current metering between 2001 and 2004 were supplied by Thomas Roy (Cascadia Biological) for the lower mainstem site M1. This site was identified by the staff gauge and VIA SAT data logger box still at the site, which is the same site used by all monitoring proponents that contributed data to this project. In reality, the conditions in the mainstem channel violate the assumptions of both methods as discussed below. Fortunately the bedrock control is very stable so the rating curve is
12 not expected to change, and close agreement among the different sources of data suggested that the data could be pooled together to derive a common rating curve. The rating curves developed as described above are shown in Figures 6 – 8. The coefficients of the curves are given in Table 3. The accuracy of the measurement can be thought of as the mean deviation between the measured flow and the rating curve for a given gauge height (Table 4). The accuracy is between 6 and 11%. The main source of error in this case is the degree to which the assumptions of the methods are violated. For salt dilution the main sources of accuracy error in this case are the presence of pools (mainstem) inflows or outflows to or from the channel in the gauging reach (tributaries).
1.25 1.20 1.15 1.10 1.05 1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0
Gauge Height (m)
M1 M2 M1 rating curve M2 rating curve
Discharge (m3/s)
Figure 6: Rating curves for M1 and M2
13
1.10 1.05 1.00 0.95 0.90 0.85
Gauge Height (m)
0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00 1.05 1.10 1.15 1.20 1.25
M1 Rating curve M1 M2 M2 rating curve
Discharge (m3/s)
Figure 7: Rating curves for T1 and T2
0.850
0.800 Rating curve Flow measurements
0.750
0.700
Gauge Height (m)
0.650
0.600
0.550
0.500
0.450
0.400
0.350
0.300 0.000
0.050
0.100
0.150
0.200
0.250
0.300
0.350
0.400
Discharge (m3/s)
Figure 8: T2 rating curve.
14 Table 3: Coefficients of rating curves:
Site M1 M2 M3 T1 T2 A 2.09 1.50 0.64 0.62 Offset 0.166 0.10 0.41 0.21 b 0.069 0.140 0.410 1.060 c 0.547 0.370 0.280 0.180 GH Asymptote 2.256 1.600 1.050 0.830 Min GH 0.206 0.136 0.410 0.213 Transition GH 0.900 0.643 0.850 0.639 B1 5.525 2.444
B
B0 -2.6894 -1.1118
B
5.648 1.800
-4.059 -1.020
Note: A, Offset, b and c are all curve fitting parameters for equations 1 and 2. GH Asymptote = A + Offset. Min GH is the minimum gauge height recorded at extreme low flow and represents the zero flow point. Transition GH is the gauge height at which the rating curve is extrapolated by a linear relationship (equations 3 and 4), and B1 and B0 are the coefficients of that relationship. Table 4: Accuracy of measurements / rating curves.
Site M1 (2003-04) (2003-04) (2003-04) (2001-02) (2001-02) (2001-02) (2005/06) (2005/06) (2005/06) RMS error (%) M2 Q (m3/s) 0.051 0.085 0.847 0.024 0.727 1.012 2.139 0.081 0.017 GH (m) 0.275 0.295 0.570 0.280 0.560 0.670 0.874 0.316 0.217 Q calc 0.066 0.089 0.737 not used 0.703 1.118 2.156 not used 0.016 % error -22.30 -4.90 14.98 3.43 -9.50 -0.78 4.01 11.1 9.44 -4.19 -5.73 6.8 -1.77 -11.31 -5.00 -13.39 9.2 -15.63 -5.81 -0.06 -1.60 8.4
0.383 0.036 0.553
0.587 0.315 0.686
0.350 0.038 0.587
RMS error (%) T1 0.040 0.886 0.465 0.015 0.613 0.882 0.807 0.570 0.041 0.999 0.489 0.017
RMS error (%) T2 0.013 0.123 0.089 0.005 0.506 0.639 0.612 0.452 0.016 0.131 0.089 0.005
RMS error (%)
15 Discussion of rating issues: The tributary flows were straight-forward to measure since this involved routine application of salt dilution. The mainstem sites M1 and M2 were very challenging. Eventually we found that salt dilution using the mass balance approach but with the salt in solution form seemed to produce the most reliable results. Although the TOR called for 5 measurements at each site to establish rating curves, with each measurement collected in triplicate to establish error, we did not achieve this at all sites. We collected 4 points on each of the tributaries T2 and T1, and 3 each on the middle mainstem site (M2) and on the lower mainstem site (M1). These numbers are the reliable measurements, unreliable points are excluded. The upper mainstem site (M3) could not be measured, only estimated. Given what I know about rating curve accuracy, I don’t believe there is a difference between 3, 4 or 5 points. For the kind of streams we are working in, the true test of rating accuracy should be based on a statistical analysis of 30+ data points across the full range of flows, something that is not possible with the existing standards. We are currently testing a prototype of a system to deliver salt to a stream automatically, thereby allowing for salt dilution gauging to be applied by a data logger at selected water levels. If such a system were to be used routinely, rating curves could be developed relatively quickly and based on a sample size large enough for the error to be determined statistically (without violating the assumptions of the methods). Until then we must live with the uncertainties in rating curve development. The standard methods involve assuming that all measurements conform to a known error (e.g., less than 5%) and forcing the rating curve to pass within 5% of each data point. Any data points that conflict with this assumption are eliminated, thereby making a sparse data set even sparser. We don’t know what level of precision is achievable in turbulent streams with irregular beds. In the absence of such knowledge it is more important to ensure that the measured flows cover the range of flows of interest to the study with a uniform distribution. The data that we have conform to the curve fitting technique that we use for all streams and cover a reasonable range of flows. I have confidence in the tributary flows. However, the lower Jordan River mainstem almost defies measurement because the flows are far too small for the size of channel. It violates the assumptions of both current metering and dilution; the bed is too irregular for the former, and the very large volume of water stored in pools relative to the flow rate interferes with the latter. We resolved this by measuring the flow at times when the pools appeared not to be connected to the main flow at specific gauging sections (these conditions were extremely difficult to isolate). The flows at M1 were validated in comparison with data supplied from 2001 – 04, and then the measured flows at M2 were validated relative to M1. We propose to use Rhodamine dye to demonstrate visually the degree of mixing under conditions similar to when the data points were measured and thereby validate the rating curves. Flow Budget: To answer questions about whether or not there are significant losses to groundwater from the channel, a flow budget was constructed by summing up the inflows from tributaries and residual areas and comparing those flows with the mainstem flows. Because all the tributaries were not measured, it was necessary to make some assumptions about the relationships between the un-gauged tributaries and the ones
16 that were gauged. Typically, west coast watersheds on Vancouver Island are characterized by steep-sided narrow canyon-like channels with very strong precipitation gradients. The change in precipitation volume and intensity with elevation is often far greater than that which could be explained by orographic processes alone. This suggests that the steep narrow topography may have a funneling effect on precipitation, but the underlying physics behind this process is not understood. In the absence of understanding of the underlying processes, it was necessary to assume that all parts of the Jordan River watershed have similar precipitation-area-elevation gradients. Therefore basin area-elevation information was used to estimate flows from un-gauged tributaries relative to gauged ones based on similarities between catchments. The flow budget was calculated for the lower half of the mainstem channel between M2 and M1 gauging stations for three time periods: the data record between December and April, and two low-flow periods, one in December and the other in August. For each period the flow from un-gauged areas was calculated based on five hypothetical reference combinations for each watershed (Table 4). To estimate the flow from T1.5 and M1resid, the flow from the reference was multiplied by the ratio of the area of the unit for which the estimate is needed over the reference area. This can be expressed as:
Q = Qref ×
A Aref
(5)
where Q is the flow estimate required for a given catchment, A is the drainage area of the catchment for which the estimate is required, Qref and Aref are the flow and area of the measured reference. The flow budget scenarios are constructed for the purpose of determining whether there is evidence of groundwater inflow to or leakage of water from the channel between the middle and lower mainstem gauges. The flow budget consists of summing the flows over a specified time interval from M2, T1, T2, and the ungauged area that includes T1.5 and the residual M1 area. The ungauged contributions must be estimated; the estimates are made using a combination of proportional area relative to a gauged area that the ungauged area is thought to resemble hydrologically. For example, T1.5 may be similar to T1, T2, or some combination of the two (Table 5). Using area-elevation information does not shed much light on which of the scenarios is correct for estimating flows from T1.5 or the residual area of M1. Using the flow record from December to the end of May, the flow budget results in a range of outcomes from a mean flow loss of 0.057 m3/s (i.e., leakage, assuming T1 as a reference for T1.5 and M1 as reference for M1residual) to a mean gain of 0.079 m3/s (i.e., groundwater inflow, assuming references T2 and M2 for T1.5 and M1residual). However, using the mid0Desember flow recession period and the summer low flow period, the budget scenarios all give similar results and all indicate a gain of water, which suggests quite clearly that there are groundwater inflows to the channel that are not accounted for by the flow budget. If drainage area is a good predictor of flow one would expect a perfectly linear relationship between area and total discharge. However this is not the case as shown in Figure 9. Using total flow for the fall – spring period the flow from T1 is higher than expected, but using only the recession period in mid-December the total flow from T1 is
17 lower than expected. For M2 the opposite is true, indicating that M2 is dominated by groundwater flow while T1 is dominated by runoff. This also indicates that elevation distribution is an important factor. Therefore it is reasonable to expect creeks with larger proportions of higher elevation (and still being rain dominated) such as Sinn Fein Creek to produce more flow per unit area than watersheds that are predominantly lowelevation. Conversely, T1 also drains more rapidly, probably due to its steeper gradient than other parts of the watershed.
Table 5: Three different budgets with different scenarios for estimating data for ungauged tributaries.
reference for T1.5, M1res Fall - Spring M1 T1 T1.5 T2 M2 M1resid Subtotal Gain / loss of flow Mean flow gain/loss (m3/s) Low flow December 14 - 19 M1 T1 T1.5 T2 M2 M1resid Subtotal Gain / loss of flow Mean flow gain/loss (m3/s) Low flow August 2006 M1 T1 T1.5 T2 M2 M1resid Subtotal Gain / loss of flow Mean flow gain/loss (m3/s) 17.95 3.392 1.872 1.315 8.224 3.149 Areas (km2) 17.95 3.392 1.872 1.315 8.224 3.149 T1, M2 15463 4879 2693 1089 4918 1883 15461 1 0.000 17.95 3.392 1.872 1.315 8.224 3.149 107172.1 12333.1 6806.5 5987.1 54737.5 20959.2 100823.4 6348.6 0.013 27707.9 0.2 0.1 0.0 1220.3 467.3 1687.9 26020.0 0.013 T2, M1 T2, M2 T1, M1 Total flow (dm3) 15463 15463 15463 4879 4879 4879 1550 1550 2693 1089 1089 1089 4918 4918 4918 2712 1883 2712 15149 314 0.022 (T1+T2)/2, (M1+M2)/2 15463 4879 2122 1089 4918 2298 15305 158 0.011 107172.1 12333.1 7664.8 5987.1 54737.5 19879 100601.8 6570.3 0.014 27707.9 0.2 0.1 0.0 1220.3 2664 3884.4 23823.5 0.011
14319 16291 1143 -828 0.079 -0.057 3 Total flow (m ) 107172.1 107172.1 107172.1 12333.1 12333.1 12333.1 8523.1 8523.1 6806.5 5987.1 5987.1 5987.1 54737.5 54737.5 54737.5 18799.3 20959.2 18799.3 100380.1 6792.0 0.014 27707.9 0.2 0.0 0.0 1220.3 4860.3 6080.8 21627.1 0.010 102540.0 4632.1 0.010 Total flow (m3) 27707.9 0.2 0.0 0.0 1220.3 467.3 1687.8 26020.1 0.013 98663.5 8508.5 0.018 27707.9 0.2 0.1 0.0 1220.3 4860.3 6081.0 21627.0 0.010
18
120000 Total Flow Dec Recession (m )
3
18000 Recesssion Dec. 14-19 Fall - Spring (168 days) Total Flow Fall-Spring (Dm ) 16000 14000
3
100000 80000
M1 M2 T1
12000 10000 8000
60000 40000
6000
T2
20000
4000 2000
0 0 5 10 15 20
0
Drainage Area (km2)
Figure 9: Total measured flows for two periods at four gauged sites as a function of drainage area. This leaves the question: which flow budget scenario is right? On one hand it could be argued that it does not matter since the flow loss or gain is an order of magnitude less than the planned flow release of 0.25 m3/s. On the other hand there is enough uncertainty in the flow budget to suggest that it would be advisable to gauge T1.5 just to be safe and continue collecting flow measurements at all sites to verify that the rating curves are correct. Progress towards primary objectives for the project: We have collected time series data on the mainstem channel and selected tributaries and demonstrated its precision and accuracy. Results that we have at this point suggest that there is no leakage from the mainstem channel between M2 and M1. However there is uncertainty in the remaining tributary flow estimates that must be resolved with further monitoring including the addition of another gauging station on tributary T1.5. Difficulties in measuring flow at the upper mainstem site M3 also need to be resolved before the potential for leakage from the upper part of the channel can be assessed. Once these details have been attended to and additional data collected, we will be in a better position to determine the accuracy of the flow estimates that led to the WUP decision and to answer the questions concerning leakage from the entire length of the lower Jordan River channel. Recommendations: Based on the discussions given above and given the proposed 6-month advancement of the planned flow release date we recommend the following:
19 1. Continue collecting hydrometric data at all sites over a similar or greater range of flows for the 2006-07 contract with Rhodamine dye application to visually verify mixing, specifically at the mainstem sites. 2. Install a weir at the outlet of the pool where M3 is located to improve the ability to measure the flow rate. 3. Install a stream gauge on T1.5 near the outlet to the Jordan River mainstem. Collect hydrometric data sufficient to develop a rating curve. 4. Collect data until January, at which time an interim report will be generated that will form the basis of recommendations in support of the proposed flow release. 5. If the above recommendations are accepted the 6-month advancement of the release date should not cause any problem. We have demonstrated the quality of the data collected thus far and the additional data should answer the remaining questions by January assuming a normal distribution of storms during the fall, such that the release can be activated in the summer of 2007 as proposed. Conclusion: The data collected since November 2005 indicate that there is groundwater inflow in the order of 0.01 m3/s that sustains low flows even in the absence of tributary inflows in late summer. Despite the extreme difficulty in measuring flow in the mainstem channel, particularly under low flow conditions, measurement precision of 5% and rating accuracy of 10% or less was achieved. This means that at low flow at M1, the flow is known to within 0.002 m3/s. Based on the above, the proposed release of 0.25 m3/s of flow from Elliot dam should result in a measurable increase in streamflow, since the inflow is in the order of 1/10 and the measurement error in the order of 1/100 of the proposed release. However, this is based on incomplete data that must be validated over the next contract period and prior to implementation of the release. References: Hudson, R. and J Fraser. 2005. Introduction to Salt Dilution Gauging for Streamflow Measurement Part IV: The Mass Balance (or Dry Injection) Method. Streamline Watershed Management Bulletin 9(1): 6 – 12. Hudson, R. (in preparation). A geomorphic approach to rating curve development. Forest Research Technical Report TR0XX (Hydrology). Coast Forest Region, Nanaimo B.C. Ministry of Environment, Lands and Parks, 1998. Manual of Standard Operating Procedures for Hydrometric Surveys in British Columbia. Prepared for the Resources Inventory (Standards) Committee, Victoria, BC ISBN0-7726-3484-X
Sit, V. 1994. Catalogue of Curves for Curve Fitting. Biometrics information handbook #4, Province of B.C., Ministry of Forests, Victoria BC.
20 Appendix A: Jordan River Water Use Plan Monitoring Program Terms of Reference Lower Jordan River Inflow Monitoring 1 Monitoring Program Rationale 1.1 Background The Jordan River CC's recommendation (BC Hydro 2002) to release a base flow was based, in part, on estimates of weighted usable rearing area (WUA). For a given section of the river (x), WUA was based both on assumed local inflow, QLocal(x), and the flow release (QBaseFlow): WUA(x) = ƒ(QLocal(x) + QBaseFlow) The combination of local inflow and the base flow selected by the CC was predicted to yield approximately 3 km of additional wetted habitat in the upper reaches. The Jordan River CC has recommended that more accurate river discharge and local inflow contributions be assessed. This information is required to confirm benefits to resident fish habitat in the Lower Jordan River that were calculated using an assumed local inflow and a 0.25m3s-1 flow release. The information will also be used to determine the actual flows in the river and if there are any sub-surface flow losses. 1.2 Management Questions The primary management questions discussed regarding the natural inflows below Elliott Dam are: 4. How accurate were the assumptions of local inflows used for WUP recommendations accurate? 5. What implications, if any, are there on the WUP recommendations based on revised inflow data? 6. What are the reasons for the differences, if any, between the monitored and assumed inflows? No real time series data for local inflow below Elliott Dam was available for the purposes of calculating the WUA during the Jordan River WUP. Total local inflow below Elliott Dam was calculated based on the proportional contribution of daily surface and tributary inflows for the next upstream catchment area, the drainage area for Elliott Headpond (BC Hydro in Draft). Similarly, contribution of flow along the linear length of the river below Elliott Dam was determined as a proportion of discrete drainage area to total drainage area. It should also be noted that the daily inflow to the Elliott Dam catchment, the reference watershed, was also not directly measured. Daily inflow to the Elliott Dam catchment was back calculated from changes in reservoir levels, spill, and turbine discharge for the entire system. The culmination of these factors, required the FTC to acknowledge the high degree of uncertainty associated with the local inflow data set. In addition, base flow releases were modeled in the WUA estimate assuming all base flows contribute to the downstream watercourse. It is possible, however, that subsurface conveyance losses in the dry section of the channel immediately below Elliott Dam may negate any benefits associated with the base flow release from Elliott Dam in providing additional wetted habitat. Following the collection of these data, if estimates of local inflow contribution are not underestimated and a base flow release is implemented, the stations will subsequently monitor the efficacy of the base flow release at both downstream sites. 1.3 Summary of Competing Hypotheses The following competing hypotheses have been proposed to address the management questions outlined above:
21 H1: Estimated instream flow overestimates monitored instream flow. H1a: Proportional watershed area calculations for runoff are not representative of actual instream flows. H2: Groundwater losses undermine ability for base flow releases to meet habitat expectations. 1.4 H2a: Estimated instream flows underestimated losses due to groundwater flow. Key Water Use Decision Affected
The key water use decision affected by the results is amount of water released. This monitoring program will determine the extent to which actual minimum flows below Elliott Dam were modeled in the Jordan WUP. The recommendation of 0.25m3s-1 release from the dam has an associated habitat connectivity benefit that will be tested before and after the base flow implementation. At the end of the 6year review period, the base flow release will be 0.25m3s-1 or less depending on the amount of water release required to meet the habitat expectation. 2 Monitoring Program Proposal 2.1 Objective and Scope The objective of this monitoring program is to assess the performance of the key WUP decision to increase flows in the lower Jordan River from leakage/local inflows to ≥ 0.25m3s-1 using instream flow measurements as the performance measure. The monitoring program will be assessed at two points of the lower Jordan River below Elliott Dam (reach 4 and reach 8 of Figure 1). The program will be in effect for six years – two years prior to the implementation of the minimum flow and four years after.
Figure 1: Lower Jordan River. River profile and approximate distances of significant reach breaks relative to the tailrace. Flow continuity stylised as line thickness.
22 2.2 Approach
This monitoring program will monitor instream flow at three locations in Lower Jordan River through the combination of water level monitoring and rating curve (flow-water surface elevation relationship) development. Opportunistic evaluations of “no-flow” regions will be provided during the monitoring during known low inflow periods, but this aspect of monitoring is to be collected in other monitoring programs for this facility. 2.3 Methods 2.3.1 Site Selection A hydrologist will choose three sites for the locations of the data collection platforms in consideration of the data requirements associated with the management questions for this monitoring program. The hydrologist should consider those sites: • • • • • • that are minimally affected by groundwater losses; that provide channel control for stability and water level monitoring; adjacent or proximal to suitable water measurement locations; with reasonable access; adequately protected from vandalism or damage from debris/inflows; and that reasonably articulate the variation in inflows through the lower Jordan River.
There are currently two sites that have been established for the purposes of WUP data collection that may be used for the purposes of WUP monitoring (one located approximately 200m upstream of the Jordan GS tailrace, and the other just upstream of Sinn Fein Creek confluence). A third site should be located proximal to the Elliot Dam outlet to monitor leakage and fish release flows. 2.3.2 Installation of Data Collection Platforms (DCP) Installation of pressure transducer, data logger and staff gauge at each site will be done according to standard practices. The data logger will be kept dry and must be accessible at high water. The installation must be stable through a range of events, and the staff gauge must be visible from a preferred vantage-point. The transducer cable will be protected, and all points of installation will be anchored to large boulders where possible. The staff gauge and transducer installation will be surveyed to a local bench-mark, and a site drawing (by hand) will be completed to illustrate the location and survey information. Photos will be taken to document the installation. The Elliot Dam installation site should be done such that impacts of large spill events will be mitigated, and access considers safety aspects associated with being located downstream of the dam. It is expected that one of the loggers will be replaced over the six-year period due to environmental conditions. The loggers will be set to collect water levels at 15-minute intervals. 2.3.3 Rating Curve Development Flow-stage relationships (rating curves) for each DCP will be gathered over a range of low flow conditions (0-5cms) at a site proximal to the DCP location. Alternatively, this data will be collected during fish release valve testing over a smaller time period. A minimum of five flow conditions will be measured to articulate the rating curve. Each measurement will be completed three (3) times to establish standard deviation. Measurements will be completed according to standard practice, and related to a staff gauge measurement. Consult the Resource Inventory Standards Committee’s hydrometric standards (1998) for detailed methodologies. Photos will be taken from an established photopoint to document the range of flow conditions observed.
23 If the error is large compared to the flow of interest (0.25m3s-1) – i.e. SD > 0.1cms – other means of measurement must be sought. Site modifications and measurement devices should be geared towards the collection of low flow information. Three rating curve periods are planned over the 6 year program. 2.3.4 Data Download Data are to be downloaded from the sites once every four months, preferably prior to inflow events, if possible. Where opportunities exist due to activities related to this program, data should be downloaded at a higher frequency. Each download will be an opportunity to ensure the data logger and staff gauge are synchronized per the installation. Data collection by the logger will be observed in the field. If the data are not synchronized, the site will be re-surveyed once the set-up is properly stabilized. 2.3.5 Reporting An annual report will be provided at the end of each fall season summarizing the instream flows at each site for November 1st to October 31st period. Graphical presentation of results (annual hydrograph, absolute maximum and minimum flows, etc.) and rating curve changes with a short description of methods and equipment used will satisfy the reporting requirements. 2.4 Interpretation of Monitoring Program Results The decision to release 0.25m3s-1 from Elliott Dam was predicated on the contribution of local inflow and an effective base flow release to provide an additional 3 km of habitat immediately downstream of Elliott Dam. The decision will also preserve flows ≥ 0.25m3s-1 for the length the Lower Jordan River. If after monitoring inflows over the 6year review period it is evident that the assumption of the habitat benefits associated with local inflows based on drainage area calculations were underestimated, the necessity to provide a 0.25m3s-1 base flow may be reduced. Habitat benefits will be assessed in terms of WUA and habitat connectivity.
Lower Jordan River Discharge: Pre Base Flow Monitoring 2 Years Install transducers and established stage discharge relationships. Monitor discharge before base flow release.
Lower Jordan River Discharge: Post Base Flow Monitoring 4 Years
Monitor discharge with 0.25 m3/s base flow release.
End of Review Period Review Monitoring Results
Satisfies WUP performance targets.
No
Do measured inflows exceed WUP modeled inflows?
Yes
Does base flow meet/ exceed expected gains?
No
Yes
Review WUP Requirements
Figure 2: Operational and Environmental Implications of Accurate Local Inflow Data
24 2.5 Schedule
Pre base flow release data will be collected in the first two field seasons following the approval of the Water Use Plan. Four years of post base flow release data will be subsequently collected pending engineering completion of the release mechanism and establishing a stable operating regime. 2.6 Estimated Budget The following budget assumes that consultants will provide the technical resources to complete this monitoring program over the 6 year period. The total costs of the program is $49.9K or an average of $8.3K/year. Costs include 15% contingency and administration for field study costs.
Number of Units Program Expense Type Rate YR4 YR5 YR6 YR1 YR2 YR3 Cost Tasks Site selection Hydrologist 700 1 $ 700 Monitoring Installation Technician 300 6 3 3 $ 3,600 Rating Curve Development Technician 300 12 12 12 $ 10,800 Data Download/Maintenance Technician 300 3 3 3 3 3 3 $ 5,400 Reporting Hydrologist 700 1 2 1 1 1 3 $ 6,300 Mileage 0.4 575 125 475 100 475 150 $ 760 Expenses DCP Equipment 3000 3 1.5 $ 13,500 Equipment Rental 100 9 0 7.5 0 7.5 0 $ 2,400 Room and Board 100 11 1.5 9 1.5 9 1.5 $ 3,350 Contingency 0.05 7000 900 5400 900 5400 900 $ 1,025 Miscellaneous Administration 0.1 7000 900 5400 900 5400 900 $ 2,050 Total (2004 $) $ 19,980 $ 2,635 $ 8,750 $ 6,425 $ 8,750 $ 3,345 $ 49,885
3
References
BC Hydro, 2002. Jordan River Water Use Plan Consultative Report. Prepared for the BC Hydro Jordan River Water Use Plan project, Burnaby, BC ISBN0-7726-4722-4 BC Hydro, in draft. Jordan River Fisheries Information Overview. Prepared for the BC Hydro Jordan River Water Use Plan project, Burnaby, BC. Ministry of Environment, Lands and Parks, 1998. Manual of Standard Operating Procedures for Hydrometric Surveys in British Columbia. Prepared for the Resources Inventory (Standards) Committee, Victoria, BC ISBN0-7726-3484-X
25 Appendix B: Example of a salt dilution measurement. This example salt dilution measurement was performed at M1 on October 05, 2006 under low flow conditions. The gauging reach contains pools that sometime interfere with salt dilution, however under these conditions, the water level is so low that the pools are either small (Figures B1 and B2) or disconnected from the flow (Figure B3). On this occasion two salt slugs of 0.7 kg each were weighed in advance and injected at two different locations: the first at the upstream end of the reach (Figure B4) and the second one at a point about 1/3 of the reach length from the upstream end (Figure B5). Because of the conditions, the salt masses had to be dissolved prior to injection in 20 litres of stream water. Injection points were cascade points where potential mixing was maximized. The resultant electrical conductivity (EC) over time graph (Figure B6) shows the effect of varying the mixing length; the first slug produced a damped, drawn-out dispersion graph while the second slug resulted in a much narrower, peaked graph. Although these two graphs look quite different the areas under them are virtually the same. The equation to calculate discharge is:
Q=
M A
(B1)
where M is the injection mass of salt and A is the area under the concentration over time graph, calculated as: A = I ∗ c = I ∗ CF ∗ EC (B2)
∑
∑
where I is the time interval between successive measurements (usually 5 seconds) c is the concentration (mg/L), CF is the temperature dependent concentration factor and EC is electrical conductivity. Note that equation B1 is dimensionally correct; for Q in m3/s, mass should be in grams. To determine the flow the following calculations were done (see spreadsheet, M1 10-0506): 1. Subtract the baseline and convert the EC to concentration of salt: at 10.5 degrees C, the concentration factor (CF) is 0.685. In Figure B6, the red graph is measurement #1 and the blue graph is #2. 2. Calculate the areas under the red and blue graphs. On the spreadsheet simply take the sum of the columns and multiply by 5 seconds. 3. calculate Q using equation B1 The flow is 0.017 m3/s.
26
Fig. B1: JOR M1 gauging reach (upper 2/3 looking upstream).
Fig. B2: JOR M1 gauging reach (lower 1/3 looking downstream).
27
Fig. B3: upper gauging reach; note pools in the right foreground are not connected to the flow.
Fig. B4: Upper injection site; overflow to left and right of the large mid-channel boulder
28
Fig.B5: lower (second) injection site into the overflow point in centre of the photo.
500.0 Electrical conductivity (μ S/cm); Concentration (mg/L) 450.0 400.0 350.0 300.0 250.0 200.0 150.0 100.0 50.0 0.0 0 500 1000 1500 2000 2500 3000 Time (seconds)
Fig. B6: EC and concentration over time for 2 successive dumps of 0.7 kg salt.
EC c1 c2