I’ve been meaning to put together a photo gallery of installations. So here it is (if the stand looks off-kilter, it’s just the photographer and not the stand). Please note that Rating Curves have not been developed for all sites. Where RCs are presented, they are the most recent RC from the Salt Portal, and may contain only a few days of data, or a few years. Where multiple rating curves have been developed over time, a more elaborate figure is presented. All sites are a bit different and we will work with you to find the best solution for your site, including finding a local supplier of tank and the best telemetry solution.

McKeand River, Near Iqaluit, Nunavut (Sep 2024)

Base Camp in Qikiqtaaluk

Another otherwordly adventure in Nunavut. This time to install two Big Freakin’ Flow (BFF) Auto-Salt Systems (ASS) for a potential hydropower project near Iqaluit, Nunavut. Each system consists of a coupled 300+1000 Litre tank with 20% Brine, for a total of 260kg of salt ready to inject. We are aiming to measure up to 500 m

Site B AutoSalt

Site A AutoSalt

This landscape was so amazing: part Lord of the Rings (the books, not the movie) and part Mars Terraforming., every photo deserved a panorama. Greens and brown and yellows, horizons punctuated with basalt blocks. So much land, so few people. We’ve only been able to process a few measurements so far, and shockingly, as little as 5 kg was seen two kms downstream and the SDIQ agreed with the injection measured only 300m downstream.

Moor House UK EA Station, (Sep 2024)

About 2 hours north of Reading in the Dales is Moor house. “Beware the Moors!” they whispered to Sherlock. This UK EA station has been notoriously difficult to get to for large flow event, recording a maximum of 8 cms in over 30 years of record. The AutoSalt recorded a full storm event up to 10 cms in the course of 1 month.

A slightly unmixed result, but the Left Bank agrees very well with historical injections.

Low Houses UK EA Station (Sept 2024)

After installing Moorhouse, we went to Low Houses in the Yorkshires. This is large flat reach, but amazingly the reach achieves complete mixing. There is evidence of hysteresis at this site (velocities increase first before the water level rises at the gauge). This is indicated by the larger Q for the same water level on the rising limb as the falling limb. Ignore the Qs of 1.0 and water levels of 0.1; they are upstream probe data.

Trail Valley Creek, Near Inuvik NorthWestTerritories (May 2024)

I helped Arnab Singh install an AutoSalt on loan from WSC in April at the Trail Valley Creek Research Station. After some hiccups, 71 measurements have been made with an average error of 5.8% for a Grade A rating. There was no creek when we installed it.

Rouziguiaou Creek in Brittany, France (April 2024)

This site was quite startling when I arrived. “You do know that salt dilution requires mixing to work, right?” The site is essentially a marshland, however there is an upstream culvert which contains all the flow. We positioned 3 AT units downstream, about 10m only, equally spaced across the channel. The real innovation here is an injection pipe with holes spaced approximately proportional to the flow beneath each flow. This is a way of “pre-mixing” the brine across the channel. Once this was done, we were getting fairly good agreement between the three probes. Even when not completely mixed, the QComp should give a good estimate of Flow.

Heltondale in Cumbria, UK (Jul-Oct 2023)

We’ve created a temporary RC based on the loose PT for the peak flood event. Imagine 6cms surging through this tiny channel

See the discrepancy in the stage, but the continuous Hydrograph: Magic

This slide compares the EA Stage-Based Hydrograph to the AQ Stage Hydrograph

There was a small hydraulic change due to the Sept 20 event at the EA Heltondale station, confirmed by EA staff.

A picturesque pastoral portrait

The Heltondale AutoSalt for University of Lancaster, care of Nick Chappell, was slow to get started (installed in May 2023) after several failed Boost converters for the PicoMag (needs 20V) and weird Teltonika firmware issues, it was off to the races in September. It captured a large event in Sep 2023 and has been working flawlessly (more-or-less) since then capturing 208 measurements with an average uncertainty of 6.5% for a Grade A.

Olen Creek in Italian Alps, Italy (Sep-Oct 2023)

I had the distinct pleasure of meeting Stefano Ferraris and tromping through the alps north of Turin for the Politecnico di Torino’s AutoSalt installation on Olen Creek. 31 measurement so far with excellent solar charging. The range of flows is currently small however. Excited to see how this unit performs over the winter.

Wedgemount Creek near Whistler, B.C. (2021-2023)

Looking downstream

Looking Upstream

Digging out the AT dataloggers in February was fun!

Sam was having fun

70318250108 4a9d170b 9d43 487c 82b5 095825ce11c5

Autosalt in April, insulated under 6ft (1Fathom) of snow

Current Rating Curve

All 5 historical rating curves

The Wedgemount Creek Intake AutoSalt is used to monitor flows downstream of the hydropower intake. The hydraulic control is very unstable, typical of high mountain streams, so the rating curve changes often. This system worked throughout the winter when it was buried in 6ft of snow, even the LoRa telemetry. The 20% brine stays fluid down until -20oC. While there have been 5 rating curves since installation in 2021, the average %Error is 5.6% with 115 SDIQs (left and right bank, so about 60 injections, unique stage-discharge pairs).

Blaney Creek in the UBC Malcolm Knapp Research Forest, B.C. Canada (Dec 2022)

Finishing up in the December night with Brian Wang.

The UBC Blaney Creek AutoSalt system is unique. It features wifi telemetry and 3 downstream monitoring stations, much like the Eskay Creek site, described below. Injections made at the upstream site can be measured at the three downstream locations, so long as they are not losing reaches (none are expected to be) since salt dilution will only measure increases in flow (increased dilution) moving downstream. So far the results are very promising, with 50 injections spanning the range from 0.183 m3/s to 10 m3/s. Because we know the distances between sites, we can also work out a channel reach Q-Velocity relationship. (TBD)

The rating curve so far at the Blaney Creek Observatory station.

The Blaney Creek Observatory rating curve in Log-Log space. It’s easier to see the range and fit at the lower end from this view.

The hydrograph and measurements shown in the KTS. The large flows are corroborated by both sensors (AT135 and AT136) as well as at the downstream sites.

Blaney Googleearth — Injections made at the Blaney Observatory with a peak over BG of only 2uS/cm and are graded A, appear at G-Bridge 32 mins to 80 mins later with a peak of 0.2uS/cm, but still Grade A.

Noosgulch River near Bella Coola, B.C. Canada (March-July 2022)

We installed the “Backpack” version of the AutoSalt system on a rock ledge, just upstream of a very deep canyon. The canyon is 1km long and emerges near the bridge, shown in the bottom/right most photo. The aluminum pipe showing the locatoin of the T-HRECS DL (EC.T Datalogger) can be seen on on the right (river-left). The flows range from 2-75 cms here throughout the year and given the long canyon, required 400g/cms to measure accurately. This means that for the 40cms measurement, 16kg was required, or 80 litres of brine at 20%. This requires careful management of the brine, which does not last very long for large flow events. We installed a secondary tank connected by a hose, as well as an upstream EC.T meter to better track the BG EC.T during an SDIQ in July 2022. This system also has a Neon satellite telemetry unit, and two EC.T probes measuring pumped brine conductivity and temperature.

There has been 92 measurements at this site so far, but the hydraulic control changes frequently. So far we are working with 3 rating curves since March 2022. This has been a challenging site, but the AutoSalt has performed well.

Krokbräcke Creek near Fjällbacka, Sweden (May 2022)

This lovely little stream features very large pools compared to the flow when we installed this in May. We installed three downstream EC.T probes at intervals of ~100m to capture a range of flow events.

Capture — For example, a single injection gets smoothed as it moves downstream from 10676 to 10680 to 10683. We don’t have more measurements from this site yet.

Krockbracke Update! A storm came through Fjällbacka on Nov 7, 2022 and we were able to measure up to 2-3cms, although there is evidence of either a) incomplete mixing or b) fouled sensors. This rating curve has had a few outliers labeled anomalous.

A storm came through the area around Nov 7, yipee!

This shows all three downstream AT units, along with the QComp group. The QComp is always in between the other individual measurements.

Carnation Creek Weir C for BCMoE near Port Renfrew BC, Canada (2022)

The T-HRECS is shown on the left bank. The second T-HRECS on the right bank is upstream.

AutoSalt at Weir C (show in the back left). This site used 260l of 9.4% brine, which should have only been 2% or 5% since we are aiming to dose at a rate of 100-200 g/cms. At this rate, there are about 1000 measurements (gah).

This curve is drawn in our Salt Portal with 1 transition at 0.42m. Probably it could use another transition at 0.54m. This is going up and down a hydrograph on a metal v-notch weir. I want to get a few more points before deciding on a transition.

The noise at the lower end of the curve is due to noise in the new flow meter for small volumes, but we know how to fix that. If we use 5% NaCl solution instead of the 9.4% currently used, we could inject more than 0.5 l of brine and get above the noise floor of the flow meter. We have a cellular modem and raspberry pi handling the transfer of the data from the AutoSalt to our Kronos Time Series online data portal now.

Boulder Creek, BC Hydro, Near Jones Lake, B.C. 2021

This system was installed on Oct 13, 2021, and achieved complete mixing at lower flows, but above 2cms the LB and RB derived Q diverged.

The RB probe needed to be upstream of the intake site.

The LB_Q is larger than the RB_Q above flows of 2cms. This is because the AQ injects on the RB, so more salt was coming down unmixed on the RB. The very large flood on Nov 11 washed away the Hydrometric station and scoured the entire channel. The AutoSalt was reinstalled upstream in November. Only the LB_Q was measured after Dec 8 because the SD Card was removed and reinserted with power on.

There are two distinct rating curves here, prior to Nov 11, 2021 on the left, and after Nov 25 on the right. Both RCs diverge above 2cms due to incomplete mixing. The client was only interested in flows below 2cms.

Nooklikonnik Creek for Nuxalk First Nation near Bella Coola, BC, Canada (2021)

Downstream Left Bank ECT Probe and Cool Big A$$ boulder.

The “BackPack” Version of AutoSalt is a 200L tank that straps to a backpack frame, but that’s where the convenience ends. I don’t envy the mighty Vincent who crawled under blowdown to get this to the intake site.

Nooklikonnik AutoSalt RC after 2 months of operation

Introducing… Kronos Time Series! This is the first image of the AutoSalt stage, converted to a hydrograph using the Rating Curve -> Discharge tool for multiple, compound Rating Curves. Interactive! Powerful! Fun!

Eskay Creek in Northern BC, Canada (2021)

Notice the new improved stand has a cage to protect the flow measurement equipment and serves as a handy place to put your coffee. The titlable top shelf allows hortizontal, vertical, or angled orientation to mount a solar panel.

The AutoSalt is installed on a smaller tributary of Tom Mackay, which then flows into Ketchum, thus allowing us to make measurements at 3 locations from a single injection. The orange box is the AutoSalt control box, the yellow is Unidata NRT Telemetry.

H9-Resulting hydrograph showing measurements from up and downstream probes. We measured up to 56 cms with only 5.5kg and 3.7km downstream. The PT from another consultant became dewatered in July 2021, However we were able to use the same EC-T sensors to derived a Water Quality Mixing Model (WQMM) and Water Quality Transit Time (WQTT) model to estimate the low flows.

In the WQTT(v) model, the Q is estimated as a function of the transit time. While this is not very accurate above 5cms at this site, we are seeing excellent agreement between the WQMM results and the WQTT results, both derived completely independently from one another.

In the WQMM, a smaller higher EC.T trib is measured and the downstream change in EC.T is used to calculate the total flow at the downstream site.

H9-Rating Curve. We can inject a higher dose than 100g to achieve a higher SNR at this site, which is 3.7km downstream.

H8 Rating Curve which has identified a strange shoulder in the rating curve. It seems to be legit, because the points before and after freshet agree.

H8 Hydrograph showing measurements, unsure of the validity of the shoulder yet.

At Eskay, there are 7 AT units logging temperature and conductivity continuously. This allows us to make Salt Dilution Instream Q (SDIQs) at 3 locations and build 3 rating curves from a single AutoSalt. Additionally, using the Water Quality Mixing Model, we can independently estimate the flow at a downstream site from a controlled outflow of significantly different EC water.

Stawamus River for WSC Highway 99 (08GA076)

This AutoSalt tank near the Sea To Sky Highway has adopted a spirit animal to protect it from harm.

The WSC Q is problematic at the lower range, but the AQAc derived Q generally agrees above 2 m3/s.

Stawamus RC from December 16 2020 to March 9, 2021

Martelltal in the Autonomous Province of South Tyrol, Italy (2020) for the TUM

Markus is stirring the pot here with 24 other international TUM students, Michael, Florentin, Brenda, and Dr Gabriele Chiogna, third from right, top row with a classic Italian hand gesture.

Michael and Florentin pose with their newly installed AutoSalt in the Italian Alps.

Downstream Right Bank Gauge site and Conductivity probe.

Finstertal in North Tirol, Austria (2020) for TUM

[rev_slider alias=”hs-finstertal”]
Michael is the head of Health & Safety and here’s he’s demonstrating the effect of a high voltage electric fence.

Creek PT and injection site

Downstream mixing reach.

MacKay Creek at Montroyal Boulevard (08GA061), North Vancouver, B.C. Canada (2020) for WSC

Lyssa blends in with her surroundings and the LB DS T-HRECS

2m Physical Distance AutoSalt Installation

AutoSalt system showing tank and injection system (left) and control box (grey, right)

The story at MacKay is not so straight forward. The mixing is no complete at higher flows. However, we can get a much better curve using WSC Stage and QComposite from the three downstream probes.

After processing, the QComposite flows agree well with the established WSC RC published on the WSC Realtime site. The Thalweg probe actually is very close to the RC.

Harvey Creek at Lion’s Bay, B.C. Canada (2019) for UBC

Dr. Steven Weijs Happy with his new AutoSalt system with LTE Modem capability, then slowly realizing he has been replaced by a stick.

Harvey Creek RC Sep 17, 2020

Winchie Creek near Tofino, B.C. Canada (2019) for Barkley Project Group

Winchie Creek Injection site looking downstream. The injection system injects into the tailrace flow, which then mixes with the mainstem flow and the T-HRECS are installed further downstream to measure the total flow in the Mainstem+Tailrace