Hyporheic Zone Restoration 3: Water Chemisty and Biology

Chinook salmon spawning on a plunge-pool structure at the Confluence Restoration Site, Thornton Creek, November, 2018. Photo by Chapin Pier, Seattle Public Utilities.

…Continued from Hyporheic Restoration 2…

The movement of water through the hyporheic zone in the new engineered streambed, and its changes in temperature, were reason enough to celebrate a successful restoration design. But these were only half of the story. This physical movement of water creates conditions for the changes in water chemistry that the hyporheic zone is supposed to perform. Is this happening? And the hyporheic zone itself is supposed to become a home for creatures ranging in size from bacteria on up to juvenile fish, living in the narrow spaces between the gravel, including creatures that would not normally be found in the degraded streambed of Thornton creek. Has this happened? Another fascinating part of this bold experiment was to inoculate the streambed with creatures from a more pristine stream, in hopes that this would speed up the process of biological recovery. As this worked? And finally, how have fish responded to this experiment?

These questions required involvement of several other teams of people, two teams studying the water chemistry, another studying the biological changes to the life forms living in the hyporheic zone itself, and finally, people studying fish use. These teams each included many people. I will only be able to introduce some of the main characters here. To learn the full details of these efforts and the complete list of those involved, I refer you to the original scientific articles, which will be listed at the end.

The chemical story

Conventional methods for measuring the chemical content of water are expensive, and not very sensitive. Each chemical that we desire to measure requires its own laboratory test, which can cost anywhere from $20 to $100 or more per sample, depending on the chemical target. This was a problem because we could only afford to measure a small number of chemicals, from a few specific locations at a few particular points in time. Guessing which chemicals to focus on was not straightforward. Because the conventional laboratory tests are not very sensitive, it seemed unlikely that we would be able to detect a change in water moving through one structure, or even through the entire restored stream segment. Some of the chemicals were likely to be present only in concentrations lower than the limits of detection, and the degree of error in measuring concentration was likely to be greater than the changes in concentration we were trying to detect. Fortunately, the project attracted the interest of other scientists, notably, researchers at the University of Washington, who had access to much more sophisticated equipment: a high-resolution mass spectrometer.

The mass spectrometer is a device capable of measuring hundreds to thousands of chemicals in each sample, in very small quantities. It is the perfect measurement instrument for the situation where there were a great many chemicals of interest, most of which were not substances commonly measured in routine water quality monitoring. There was also little information on which of the thousands of chemicals might be the best indicator of how well the hyporheic zone was working. Which ones would the streambed absorb best? Whish ones would be most likely to detect a difference?

Meanwhile, other research was only just beginning to identify what the really important chemicals were, from the point of view of effects on fish. This requires a short digression.

Early Spawner Mortality

For years, fisheries biologists had been noticing something very disturbing in urban streams, like Thornton Creek. Adult female Coho salmon, returning from spending their lives in the ocean, would enter the creek during the first high water from autumn storms, to spawn. They must find a suitable gravel-bedded site, often a site that had been restored at great cost, to construct a nest and lay their eggs. But before they could spawn, many of these females were dying. The syndrome came to be known as early spawner mortality. Scientists suspected that toxins in the water might be involved. But of what nature? And from where?

A team headed by Blake Feist (National Marine Fisheries Service, Seattle) first determined that the creeks where most of these females were dying were in watersheds that had the highest road densities. And not just any roads, but heavily-travelled roads, arterials. Narrowing the problem down further was very difficult. Eventually, another science team which included Julann Spromberg succeeded in collecting pure road runoff water from a nearby major highway, rushing it to the lab, and exposing adult coho salmon to this water in a controlled experiment. Salmon exposed to the road runoff while it was fresh died. Salmon exposed to an artificial solution tainted with petroleum-derived chemicals and heavy metals known to be present in urban runoff did not die. So the “smoking gun” was fresh road runoff, but the chemicals in road runoff that they thought were the culprits, turned out to be the wrong ones.

Now back to the high-resolution mass spectroscopy. Another team, which included Katherine Peter and Ed Kolodziej at the University of Washington, used high resolution mass spectroscopy to compare the chemical features of water taken from several creeks where early spawner mortality was happening, with the chemical features of pure road runoff as well as several artificially-created mixtures of water with things that automobiles release onto the road. These mixtures included fine particles of tire rubber, antifreeze, lubricating oil, transmission fluid, power steering fluid, and windshield washer fluid. By “chemical features” we mean groups of chemical compounds that are similar in their molecular size and properties, and so show up as “features” in the complicated spectrum produced by the analysis equipment. There were hundreds of these chemical features identified. By doing a statistical comparison of pure road runoff with creek water sampled during early spawner mortality, the scientists identified a “chemical signature,” that is a set of 57 chemical compounds of nine different types that were associated with early spawner mortality. And of the artificial mixtures, the one that most closely resembled this chemical signature was the solution made from fine rubber particles derived from tire wear.

More specifically, two groups of chemicals from tire wear seemed to be implicated as the most important sources of toxicity. “DPG” (short for Diphenylguanidine), which is used as a “vulcanization accelerator” in tire manufacture, showed up as important. But the most important seemed to be a group of chemicals which had not received previous attention by scientists studying environmental pollution, the “MMM” group (short for (Methoxymethyl)melamine compounds), especially ”HMMM” (short for hexa(methoxymethyl)melamine). HMMM is used as a curing agent and cross-linker in manufacture of resin-based plastics, and as an adhesion promoter in rubber products (like tires). Concentrations of HMMM as low as 10 ng/L (that’s 0.000000001 grams per liter) were associated with water in which Coho salmon were dying. During storm events at Thornton Creek, when migrating salmon would be arriving to spawn, HMMM concentrations are 30 to 200 ng/L, more than enough to make the water toxic.

Story Gets Interesting

But let’s get back to Thornton Creek. Skuyler Herzog, a researcher from Colorado School of Mines, who was associated with the team doing the mass spectrometry, injected ordinary food coloring into one of the shallow piezometer wells previously used to measure the vertical water movement, just upstream from one of the plunge-pool structures. He was able to detect where the water emerged from the streambed, downstream of the plunge pool, at two locations. Here, he installed seepage meters, which is a device that captures all of the water emerging from the streambed, so that it can be collected for analysis. Then, using a bromide solution as a tracer, he was able to measure the time that the water spent moving through the hyporheic zone (called the residence time). This turned out to be fairly short, 30 minutes for one of the flow paths (about 4.4 meters, or 14 feet long) and 40 minutes for the other flow path (5 meters, or 16 feet long).

During late autumn storms, the residence time increased in both of these flow paths. This was expected. During a storm, the high water “drowns out” the vertical step of the structure, reducing the pressure difference that drives the water into the streambed. Although this may decrease the fraction of surface water circulating into the hyporheic zone, longer residence time means greater time for chemical transformations or absorption to occur.

And when the water samples were tested with the high-resolution mass spectrometer, the results were amazing. During these autumn storms, the concentration of HMMM was reduced by 35-40% as the water passed through the longer of the two flow paths under this one plunge pool structure! And for DPG, the removal rate was over 90%. So, not only could they detect a difference, after passage through the hyporheic zone beneath a single structure, this difference was significantly large, and involved chemicals directly implicated in early spawner mortality.

How could this be? Most likely, the chemicals are not being broken down, but rather are being absorbed by the massive biofilm surrounding the grains of sand and gravel making up the hyporheic zone. This is the only way to explain such rapid decreases in concentration. Fortunately, since the biofilm is a living community of microorganisms, its capacity to absorb chemicals out of the water is continually being renewed, and unlikely to become saturated over time.

The “Smoking Gun”

With the chemical toxins responsible for early spawner mortality narrowed down to these broad groups of chemicals from tire-wear particle, there was still a lot of work to do. Which chemical was it? Was it one chemical, or many?

Using sophisticated laboratory fractionation and extraction techniques, a team of University of Washington researchers that included Zhenyu Tian (and many others) systematically split the chemical content of the tire-wear-particle solution into ever smaller numbers of chemicals, carefully testing for Coho toxicity at each stage. After 5 different fractionation runs, a solution containing 2216 chemical groups was simplified down to a solution of only 4 such groups. A final extraction process resulted in a (still toxic) pinkish-purplish substance. The only trouble was, this chemical didn’t appear in any of the literature on tire rubber or environmental toxicity.

A breakthrough came when the researchers examined the ways that complex molecules get modified when out in the environment. It turned out that one of the chemicals eliminated earlier in the extraction process (a chemical called 6PPD) could easily become transformed into this toxic substance, a complex molecule called 6PPD-quinone. At last - the “smoking gun!” The murder mystery was solved!

The molecule causing early spawner mortality

The Biological Story

Yet another team of researchers, which included Sarah Morley and Linda Rhodes from the National Marine Fisheries Service in Seattle, has been busy putting together the biological story of this grand experiment. The focus of this story is the community of microorganisms, and the insects, crustaceans, worms, and other “macro invertebrates” that live in the hyporheic zone.

To set the stage for this, there is one more aspect to the grand experiment that needs to be mentioned: inoculation of the streambed with macro invertebrates from the Cedar River. The Cedar River, which flows into the Lake Washington southeast of Thornton Creek, is part of the lands managed by the City of Seattle for drinking water. It is located in the forested foothills of the Cascade Mountains, and is closed to public entry. As such, it has very pristine water quality, and intact natural communities of streambed organisms.

The researchers took cylindrical baskets filled with clean gravel, and laid these on the streambed in the Cedar River for several months, allowing the gravel to become colonized by many of the macro invertebrates present there, many if not most of which had long disappeared from Thornton creek. Then, after the restoration work was done, these now-colonized nets were placed in specially designed wells in the streambed of Thornton creek, extending down into the hyporheic zone. The idea was that the restored streambed might now be able to support populations of these long-vanished creatures. At any rate, it would jumpstart the recolonization of the sterile streambed and hyporheic zone, much in the way that spreading seeds or planting trees helps to recolonize bare soil after it has been disturbed by fire or man.

The researchers also installed special, shallow piezometer wells to allow them to sample water directly out of the thin veneer of alluvial gravel and sand of the unrestored creek channel upstream, and also the hyporheic zone of the restored streambed. These water samples were analyzed for macro invertebrates and microbial life. In addition, they measured some of the chemicals studied in conventional water quality analysis, including dissolved nutrient content (nitrogen and phosphorus), dissolved organic carbon, and total organic and inorganic matter.

As of this writing, not all the results are in, and so the story is complex. The engineered streambed with its enhanced hyporheic zone has a higher microbial metabolic activity than the unrestored stream. This means that nutrients are being taken up, and dissolved carbon dioxide converted into organic carbon at a greater rate than before restoration. The new restored streambed has greater invertebrate density and “taxa richness.” This means that there are more macro invertebrates than before, and a greater variety of different kinds. There are changes in composition of the population of microbial and invertebrate organisms, which often indicates changes to the food web.

The inoculation of the streambed with Cedar River gravel has also had complex results. It has resulted in small, short-duration changes in composition of the microbial population and interrelationships between types. There seem to have been no significant changes in density or composition of the invertebrate population caused by the inoculation experiment. However, four “new” types of invertebrates have been found in the inoculated section of the stream.

Fortunately, these researchers intend to continue this monitoring into the future. It is difficult to draw conclusions from only a few years of monitoring, as some of the changes may become evident only over longer time spans.

Overall Assessment: Was it worth it?

Scientists can easily get caught up in the individual details of the things they are measuring. It is rare that a single measurement will definitively answer the big questions, the questions of concern to government managers in charge of water quality, taxpayers concerned about wise investment of public funds, and the curious public, who just wants to know: did it work?

In the case of the grand experiment at Thornton Creek, when we consider the teams, with all of their measurements techniques and results, it is fair to say that the experiment did work. The engineered streambed performs well physically, in terms of enhanced flow of water in and out of the hyporheic zone, enhanced hyporheic zone size, greater distribution across the streambed of diverse flow rates and temperature, and most importantly, demonstrated overall water temperature reduction. The new streambed with its special structures performs better than anybody expected in terms of water chemistry, reducing the concentrations of key pollutants that have been associated with tire-wear particles, now implicated in early spawner mortality of Coho salmon. And the new streambed seems to be on the road to biological recovery of the diversity of microbial and invertebrate life.

You will notice one glaring omission in this, and the previous two articles describing various scientific studies being done. What effect does this had on the fish?

There are ongoing measurements of fish populations in Thornton creek. The difficulty with using fish population response as a way to gauge success is that fish populations are influenced by too many other things (a scientist would say “confounding factors”) for us to know if the change we observe is due to the restoration or to something else. Fish populations respond on a regional basis to climate, ocean conditions and fish harvest, and these can mask the effects of restoration. Fish will always quickly move into a restored stream segment from other parts of the watershed, making it difficult or impossible to see if the restoration has made a difference on a watershed scale, the only scale where things really matter. But that said, with more years of fish population data, it may eventually be possible to say that this restoration project affected them.

But meanwhile, a spectacular affirmation of the potential effectiveness of this restoration experiment happened in the October of 2018, four years after completion of construction work. At the downstream restoration site, called the Forks Confluence site, a pair of Chinook salmon were seen spawning right on top of one of the plunge pool structures. The photograph that caught them in the act revealed a clean- looking, loose gravel streambed, exactly the type of place that salmon are bred to spawn in. The photograph made headlines. Chinook salmon had not been seen spawning in Thornton Creek for a great many years. All of us involved in this grand experiment found reason to rejoice.

References.

Those of you interested in further reading about the results of this restoration can find it in the following scientific publications.

Blake Feist and others, 2017. Roads to ruin: conservation threats to a sentinel species across an urban gradient. Ecological Applications, 0(0), 2017, pp. 1–15

Julann Spromberg and others, 2016. Coho salmon spawner mortality in western US urban watersheds: bioinfiltration prevents lethal storm water impacts. Journal of Applied Ecology 2016, 53, pp. 398–407

Sarah Morley and others, 2019. Biological Effects of Hyporheic Zone Restoration in an Urban Stream. Salish Sea Ecosystem Conference (2018 : Seattle, Wash.)

Sarah Morley and others, 2021. Invertebrate and microbial response to hyporheic restoration of an urban stream. Water 2021, 13, 481.

Katherine Peter and others, 2019. Evaluating emerging organic contaminant removal in an engineered hyporheic zone using high-resolution mass spectrometry. Water Research 150 (2019), pp. 140-152

Katherine Peter and others, 2018. Using High-Resolution Mass Spectrometry to Identify Organic Contaminants Linked to Urban Stormwater Mortality Syndrome in Coho Salmon. Environmental Science & Technology 2018 52 (18), 10317-10327

Zhenyu Tian and others, 2020. A ubiquitous tire-derived chemical induces acute mortalitty in coho salmon. Science, 10.1126/science.abd6951 (2020)