Establishing Correlations between Upland Forest Management Practices and Economic Consequences of Stream Turbidity in Municipal Supply Watersheds

Progress Summary: The focus of this study consists of measuring the downstream costs of increased sedimentation and determining the extent to which sediment stems from land and reservoir management activities in the North Santiam watershed, a sub-basin within western Oregon's Willamette River basin. Three tasks are being undertaken separately by teams of researchers, identified as A (Hulse),B (Whitelaw), and C (Grant) below.
1. Current data do not allow conclusive determinations regarding key project questions. To address this problem, the Hulse research team developed a geographically referenced database of natural and cultural factors relevant to project objectives. Among other uses, these have contributed to a multiagency/multidisciplinary team of professionals operating as the "North Santiam River Sediment Study Group." This group has implemented expanded monitoring efforts in the North Santiam watershed aimed at better understanding and managing detrimental effects of turbidity on municipal water supplies. The results from economic and geomorphic investigations of the research are being integrated with these geospatial data to formulate a synthetic understanding of the role of land and water management in producing turbidity and the tradeoffs to be confronted in minimizing turbidity in the future. There currently are no data to report on this task.
2. During the past year, Whitelaw, et al., have concentrated on working with our collaborators from the University of Oregon and the Oregon State University's Forest Sciences Laboratory to develop a conceptual model summarizing the central lessons offered by the Salem-Santiam case study and explaining the interactions among a municipal water utility, land managers in the watershed providing surface water for the utility's use, and the operators of dams that affect the timing and quantity of water flows. At a meeting sponsored by the U.S. Department of Agriculture Forest Service's Pacific Northwest Research Station, regarding its Water Initiative, we discussed our findings about the impacts of the 1996-97 floods on the city of Salem's water utility and water customers, and the potential implications for the managers of national-forest lands in watersheds providing water to municipal utilities. We estimated the percentage of streamflows originating, by month, on national forest lands in the North Santiam sub-basin (and other sub-basins of the Willamette River basin). Our results are summarized in a working paper, "Estimating streamflow from national forests in the Willamette River basin, Oregon," by Ernie Niemi, Michelle Gall, Matt Sayre, and Ed Whitelaw. There currently are no data to report on this task.
3. This past year's work has focused on finishing the field work, analysis, and writeup of the geomorphology section. The summer field season was directed at a stratified sampling of sediment from landforms distributed throughout the North Santiam watershed and resulted in acquisition of over 100 sediment samples. We also developed and tested a new technique for determining turbidity production from each landform. Sediment is resuspended in water and turbidity is measured over a 5-day period; the resulting time versus turbidity curve yields a systematic way of characterizing the persistent turbidity production from the sampled landform. We currently are analyzing the samples using this technique. At the same time, we are using X-ray diffraction techniques to identify the clay mineralogy of the individual samples. We will be able to directly link landforms with both turbidity potential and clay mineralogy. Because the landforms were sampled to represent distinctive sediment transport processes active in the basin (i.e., landslides, earthflows, fluvial deposition), this analysis will rigorously identify which processes are contributing turbidity-causing sediments to the North Santiam River. Using known relations between land management and process rates, we can link land management activities to turbidity production, which was the goal of this section of the project.

Results from this analysis are being written up as a Masters thesis for R. Ulrich, with a projected completion date of spring 2001. Two technical papers are planned from this thesis: (1) an examination of process/landform/turbidity relationships for the North Santiam River; and (2) a technique paper describing this new approach for assaying turbidity. We anticipate that both papers should be ready for submission by summer 2001. Development of the turbidity assay technique took longer than we anticipated, because it required careful experimentation to determine the temperature and concentration ranges necessary to get good results. Other than that, we have not encountered significant difficulties.

Preliminary Data Results: Data still are being analyzed, but we have families of turbidity versus time curves for our sediment samples that clearly demonstrate that those samples derived from smectite-rich clays have much slower settling times, hence higher turbidities than samples from nonsmectite clays. We expect that the clay mineralogy will further confirm these results.

Future Activities: In the time remaining in this study, Hulse, et al., will finalize the integration of the geospatial data and economic and geomorphic components of the project into a set of policy and land/water management recommendations for managing tradeoffs in municipal water supply watersheds using Salem, Oregon, as a case study. The project team has been meeting monthly since June 1999 to complete this synthesis.

Whitelaw, et al., will complete the final version of "The city of Salem and the North Santiam: the economic consequences of waterborne soil." We also will continue to meet with other researchers, particularly David Hulse and Gordon Grant, to coordinate the next phases of our research projects.

Grant, et al., will finish data analysis, writeup, and publication of results, including a group synthesis paper.

Journal Article:

Hulse DW, Ribe R. Land conversion and the production of wealth. Ecological Applications 2000;10(3):679-682.

Supplemental Keywords: watersheds, sediments, Pacific northwest.

The objectives of this project were to: (1) prepare an analytical framework for estimating the downstream costs from increased sedimentation caused by land and reservoir management activities; (2) employ this framework to improve understanding of the sediment costs incurred by the City of Salem, Oregon, and its water users from such activities in the Santiam watershed; and (3) coordinate with key stakeholders to identify alternatives for managing these costs.

Summary/Accomplishments:

Salem, OR, a city of 125,000 people, has long relied on the North Santiam River, flowing out of the Cascade Range into the Willamette Valley, as a municipal source of nearly pristine water (Figure 1). From February 5-9, 1996, northwestern Oregon was inundated by a series of intense storms. These storms, originating in the subtropics, brought a combination of record-breaking rain and warm temperatures. River flood stages in the first week of February 1996, were comparable in magnitude to the December 1964 flood, the largest in Oregon, since flood control reservoirs were built in the 1940s and 1950s. With the heavy rainfall and flooding, the Army Corps of Engineers flood control reservoirs and the North Santiam River received large amounts of sediment-laden water.

The February 1996 storm produced extremely high turbidity throughout the North Santiam River during and following the storm event. Turbidity readings as high as 140 nephelometric turbidity units (ntu) were measured at the City of Salem's Geren Island water treatment plant intake, which is designed to treat water at less than 10 ntu. Because the river water is normally of extremely high quality, the city uses a slow-sand filtration system that can treat pathogens and moderate turbidity and has low operating costs relative to conventional filtration systems. High turbidity, however, may damage the slow-sand filters.

The turbidity experienced in February 1996, overwhelmed the filtration capability of the city's water-treatment plant, forcing the city to close the plant for 8 days. After initiating unprecedented measures to curtail customer water use, and after building emergency pretreatment facilities, the city reengaged the treatment plant before serious water shortages materialized. However, turbidity exceeded drinking water standards set by the Environmental Protection Agency (EPA) through mid-July, and levels of turbidity remained unusually high for 5 months thereafter.

The high turbidity triggered diverse economic consequences. Damage from clogging of the filters, plus other short-run costs to the city's water utility, totaled $1.4 million (Table 1). Additional, but undocumented costs of $20-$45,000 occurred as political leaders and staff were diverted from normal activities. Some industrial water customers incurred costs of $2-3 million. Lost revenues were incurred primarily because they responded to the city's request to curtail operations during the crisis. Workers at the plants experienced a loss of $56,000 in wages during the temporary layoffs. Although the city issued an alert when the treatment plant was closed, asking all residential and commercial customers to limit non-essential uses, the overall impact apparently was no more than an inconvenience, and the alert may have triggered some water hoarding.

Table 1. Short-Run Damage to the City's Water Utility (1996 Dollars)

Category	Damage
Damage to Filter #1	$1,000,000
Cost of alternative water supplies	$204,467
Installation and operation of emergency pretreatment system	$184,320
Repairs and cleanup	$4,755
Lost revenue associated with forgone sales	$17,000
Diversion of managerial attention	Unquantified
TOTAL	$1,410,542

Source: ECONorthwest with information from the City of Salem.

Table 2. Long-Run Costs (estimated) to the City's Water Utility (1996 Dollars)

Category	Costs
Pretreatment facility	$1,200,000
Additional monitoring	$300,000
Water treatment	$56,000
New well for industrial supply	$80,000
TOTAL	$1,636,000

Source: City of Salem.

The city, at a long-run cost of $1.6 million, subsequently increased its ability to cope with future pulses of high turbidity by expanding an early-warning monitoring system, increasing its treatment capacity, securing back-up supplies from other sources, and increasing storage of treated water (see Table 2). Industrial customers also apparently incurred costs of about $80,000 to reduce the risk of future disruptions by securing back-up supplies and increasing onsite filtration capabilities. Ironically, however, this unprecedented occurrence of high turbidity clarified the value of having a watershed that usually delivers extraordinarily pure water; on average, it allows the city to avoid $2-4 million in annual treatment costs.

Figure 1. North Santiam Watershed and surrounding area

Click on image for larger version

In the weeks and months following the February floods, controversy raged over the causes of the water supply problem. The media reported that much of the sediment that caused problems for the Salem water treatment facility was derived from logging activities in the basin's headwaters. This reflected a widely held view of a direct causal link, such as shown in Figure 2a, between land use (especially forestry) and sediment production and delivery to streams, causing downstream turbidity. Despite the intrinsic appeal of such a simple syllogism, this study confirms a more complex relationship among land use activities, sediment production, and transport, and turbidity. In a real sense, the flood's muddy waters revealed the complex and often contradictory web of management objectives among the agencies and parties responsible for water management in the Santiam. It specifically highlighted how management for one narrow set of objectives might exacerbate problems in another sector. For example, decades of logging in the North Santiam Basin targeted the most unstable piece of ground, thereby potentially exacerbating production of sediment causing persistent turbidity during storms. The operation of dams for flood control captures and prolongs the release of persistent turbidity downstream, causing problems for municipal water users. Relying on the normal behavior of a watershed to produce clean water under all circumstances exposes societal vulnerability to inherent geological hazards and their interactions with land and water use decisions. Trade-offs involved in reducing persistent turbidity turned out to be much more complex in space, in resource management decisions, and in time than was hypothesized. Some background on causes of turbidity in water is useful. Turbidity is a measure of the clarity of water in relation to its concentration of light-scattering material. Turbidity in water is typically the result of dissolved and suspended fine organic and inorganic particulates, with the inorganic fraction primarily made up of clays.

In the western Cascade mountains of Oregon, this and other studies, following both the 1996 flood and earlier floods, consistently show that water exhibiting persistent turbidity contains high concentrations of a type of clay termed smectite, along with lesser amounts of other amorphous clays. Smectitic clays are widely but not uniformly distributed throughout the western Cascades. Using x-ray diffraction techniques and clay mineralogy, the source of persistent turbidity can be tied to various landforms and soils that have high quantities of smectite. This relationship among landforms, clay mineralogy, and persistent turbidity can be represented as a triangle (Figure 2b), with each leg defining a causal linkage, as indicated by the arrows, and supported by independent data. In this relationship, land use is not the cause but a contributing factor through its effect on increasing the rate of native erosion processes.

Figure 2. Two conceptual models of the sources of persistent turbidity

A more detailed look at the relationships among landforms, turbidity, and clay mineralogy reveals that earthflows overwhelmingly are the dominant source of smectitic clays producing persistent turbidity in the Santiam Basin. Samples from diverse landforms, including stream terraces, landslide deposits, earthflows, and glacial deposits show that smectitic clays comprise between 70 and 90 percent of the clay fraction from earthflows. We found smectite concentrated in only one other landform, soils derived from a debris flow that initiated above an earthflow and traveled through it en route to a stream below. Debris flows traveling through earthflow complexes may incorporate smectitic clays, and thereby provide another source of persistent turbidity. Another potential source for smectite are terraces immediately downstream from earthflow complexes that store smectite-rich sediments derived from the earthflows upstream. Although our limited sampling did not reveal smectite in such terraces and thus did not reveal the extent to which such terraces represent long-term sources of smectite, to be conservative, we included them in subsequent analyses.

Forest Land Use Effects on Sources of Persistent Turbidity

Although our analysis revealed that weathered volcanic landscapes, such as the North Santiam watershed, naturally produce clays that cause persistent turbidity, and that upstream land manager's role in causing sedimentation was less clear than originally hypothesized, we considered how forest land use activities, including logging and road construction, might accelerate or increase natural production rates of smectite and related clays. Our focus was on land use activities that accelerate sediment production or delivery processes in smectite-rich areas of the landscape (i.e., earthflows). Land use activities that we hypothesized as likely to increase sediment production or delivery from earthflows, listed in rough order of decreasing effect, include:
1. Forest harvest directly on earthflows themselves, which can accelerate erosion due to compaction, soil disturbance, or changing onsite hydrology resulting in accelerated earthflow movement.
2. Harvest in steep unstable areas prone to landsliding upstream from earthflows, where tree harvest can reduce root strength, leading to landslides and debris flows that pass through earthflows.
3. Road construction resulting in road-stream crossings upstream of earthflows; such crossings or nodes have a much higher likelihood of failure as slides or debris flows during storms.
4. Forest harvest and roads on terraces located immediately downstream of earthflows; soil compaction, disruption, and road failure in these sites might deliver smectite-rich stored sediments into stream channels.

We conducted a geographical information systems (GIS) analysis to evaluate the spatial distribution of sites where land use activities might potentially increase sediment production leading to increased persistent turbidity during storms. The maps from this analysis reveal several important conclusions about the distribution and potential land use impacts on sources of persistent turbidity within the North Santiam watershed. First, the primary sources of smectitic clays are not uniformly distributed over the basin, but are concentrated in certain areas, some of which are more active than others. In particular, most active earthflows are concentrated in the Blowout Creek sub-watershed, with lesser amounts in the Breitenbush and main North Santiam watershed. No earthflows are found within the upper North Santiam and wilderness areas that make up the eastern third of the basin. The reasons for this highly patchy distribution of earthflows are not fully understood, but likely result from the age, composition, and weathering history of the rocks in the western Cascade volcanic pile. The younger rocks making up the High Cascade region have not had sufficient time to weather to produce the high clay concentrations necessary for earthflow evolution.

Focusing in on the Blowout Creek sub-watershed reveals relationships among active earthflows, streams, and landslide trigger sites. Steep unstable zones surround many earthflows; some of these steeper areas are likely headscarps and bluffs formed as earthflows flowed downhill, leaving harder rock behind. Instability leading to landslides in these areas occurs as the underlying exposed rock along the headscarp is eroded and sapped. Most of these unstable headscarp areas are not likely to fail as true debris flows, because they typically do not concentrate water.

A more likely source of debris flows that can travel through earthflow complexes entraining smectite are steep unstable headwater areas and road-stream crossings upstream of earthflows. Not all road crossings are potential failure sites; hillslope position (i.e., upper, middle, or lower), size of stream, type, age, and dimensions of culvert are all factors determining failure potential during storms. Further, GIS analysis including these and other factors could potentially target the highest risk sites.

On the other hand, a much higher number of road crossings that could be potential trigger sites are present if both active and dormant earthflows are considered. The large areas in these older complexes coupled with dense road and stream networks mean that many more road crossings are present above old, now dormant, earthflows. A much closer look at the potential for these dormant earthflows contributing smectite should be a component of any management strategy intended to reduce land use effects on turbidity.

Conclusions:

With respect to forest and reservoir management within municipal watersheds, care must be taken to ensure that human activities superimposed on intrinsic geological conditions do not exacerbate existing erosion processes or introduce new ones. Specifically, land and water managers should recognize how watershed processes are coupled in both time and space. In the North Santiam, decades may elapse between flood events capable of producing extreme turbidity. During those decades, the combination of land use and geomorphic processes may progressively increase downstream risk. There is a nonuniform spatial distribution of landforms producing turbidity within a watershed. Different types of landforms produce different types of turbidity, as landforms and geological processes interact with each other. Consequently, an understanding of the spatial relationships and change over time among sediment production and transport processes is essential to guide land and water use and restoration activities.

More generally, human systems, including institutions, which develop with a narrow sense of time and space, become more vulnerable. The lessons of the North Santiam show that cogent watershed planning must recognize time and space scales, tradeoffs, and the imperative that what you do, and when and where you do it matter, in land and water management. A proper, more sophisticated accounting of watersheds in space and time gives more options the next time a crisis occurs. Such an accounting will provide for the correct association of public concern with those land and water management decisions that have a credible cause/effect relationship on water quality problems. New directions for the watershed that have emerged since the 1996 flood include: an early warning monitoring scheme, the prospect of higher resolution spatial planning for the U.S. Forest Service, and a broadened multi-agency dialog regarding water management; as well as a heightened public awareness of the fact that all urban residents of the Willamette River Basin live downstream from earthflows, forests, and reservoirs.
Our investigations of the 1996 flood brought to light additional, previously unseen or underappreciated interdependencies among the three agencies' (U.S. Forest Service, U.S. Army Corps of Engineers, City of Salem Public Works) use of the basin and clarified the spatial and temporal dimensions of intrinsic sediment production and turbidity, and the ways human activities may exacerbate turbidity. Trade-offs involved in reducing persistent turbidity turned out to be much more complex in space, time, and resource management decisions of institutions than was previously assumed. This complexity led, in the Santiam, to increased cooperation among the involved agencies for addressing research and monitoring, as well as land and water use decisions. In addition, the hazards presented by muddy waters will require new approaches on the part of each of the agencies involved. Our interactions with these and other stakeholders identified 10 key issues, options, and questions relevant to future land and water management in the Santiam watershed:

1. Hazard profile of increased sediment from forest lands in the basin increases with time from last major storm event.

2. Earthflows need to be viewed within their geographic context:
a. Are they downstream of potential debris flow initiation sites?
b. Do they deliver directly to stream channels?

3. Roads may interact with earthflows to change movement patterns; a sound management principle would be to not increase supply of water to earthflows via roads and perhaps reduce water supply via road/culvert re-engineering or removal.

4. If climate changes, hazard profile for persistent sediment delivery downstream may change as well (e.g., snow changing to rain).

5. Dams should be re-engineered to accommodate outflow from different pool elevations out of concern for turbidity and not solely stream temperature.

6. Flood protection equals occasional high turbidity.

7. Influence others upstream whose actions effect water quality.

8. Substitute alternative water supplies.

9. Substitute for watershed services.

10. Manage water demand.

The 1996 Salem/Santiam flood offers insights into several important watershed-management issues emerging across North America and elsewhere, with special emphasis on three domains of issues: economic, geologic, and watershed planning domains. In particular, this case study demonstrates the potential consequences of relying on a watershed to provide high-quality source water for municipal-industrial use, and the economic and institutional adjustments that must be confronted as the demand for high-quality source water rises relative to the demands for timber, flood control, and other goods and services derived from the watershed. This event indicates that the actions of land managers, river managers, and municipal water utilities are interconnected in ways not previously anticipated. Furthermore, the importance of this flood's impacts on turbidity, rather than on more conventional concerns about flood inundation, highlights the necessity for watershed managers to have a deeper understanding of a watershed's geology, looking beyond topography's impacts on the hydrograph to see how soil composition and geological processes, such as deep earth movements, interact with human activities and infrequent precipitation events influencing water quality. The 1996 Salem/Santiam flood also offers lessons for capitalizing on floods and flood-response planning to shape long-run, watershed management strategies that explicitly recognize these geological characteristics and their impacts on the supply of high-quality source water and other goods and services.

Dissertation/Thesis:

Niemi E, Whitelaw E. The economic importance of source of clean water for a municipal utility: Salem, Oregon and the North Santiam River. M.S. thesis, in preparation.

Journal Articles:

Grant G. Dam removal: panacea or Pandora for rivers? Hydrological Processes 2001;15(8):1531-1532.

Hulse D, Ribe R. Land conversion and the production of wealth. Ecological Applications 2000;10(3):679-682.

Grant GE, Glassman R, Hulse D, Niemi E. Muddy waters: linking forest land use, sediment transport, and reservoir operation to municipal water quality in the North Santiam watershed, Oregon. EOS Transactions American Geophysical Union, Boston, MA, 1998;79:S147.

Presentations:

Niemi E. Economic transitions & natural resource managers. Presented at the Annual Meeting of the American Fisheries Society, Phoenix, AZ, August 20, 2001.

Niemi E, Gall M, Sayre M, Whitelaw E. Estimating streamflow from National Forests in the Willamette River Basin, Oregon. Presented at ECONorthwest, Eugene, OR, 2000.

Grant GE, Glassman R, Hulse D, Niemi E. Muddy waters: linking forest land use, sediment transport, and reservoir operation to municipal water quality in the North Santiam watershed, Oregon. Presented at the American Geophysical Union Spring Meeting, Boston, MA, May 1998.

Grant GE. Results of North Santiam turbidity study. Presented at the Muddy Waters Symposium for Forest Service and BLM Hydrology and Soils Resource Managers, Oregon, May 24, 2000.

Niemi E, Whitelaw E, Gall M. The economic consequences of waterborne soil: Salem and the North Santiam River. Presented at ECONorthwest, Eugene, OR, 1999.

Niemi E, Whitelaw E. The economic importance of source of clean water for a municipal utility: Salem, Oregon and the North Santiam River. Presented at ECONorthwest, Eugene, OR, 2000.

Dole D, Niemi E. The future of municipal water resources in the Willamette River Basin: a basin-level analysis. Presented at ECONorthwest, Eugene, OR, 2000.