On trial in Clackamas County24 July 2009
Geographic information systems have been utilized to model and identify micro hydro sites in Clackamas County, Oregon, US. Jed Jorgensen of the Energy Trust of Oregon provides details and discusses the results of this trial study
In the US, extensive research reveals considerable development potential for small, mini, and micro hydro power resources. While broad research has been conducted nationally, resulting in tools appropriate for policymakers and larger hydro power developers, there has been less focus on site identification based on current property ownership at regional and local scales.
Energy Trust of Oregon is an independent nonprofit organization dedicated to energy efficiency and renewable energy development. To better understand the hydro resources available in Oregon, Energy Trust commissioned a geographic information system (GIS) study to determine the quantity and size of potential run-of-river projects within a sample county. Energy Trust wished to analyze this data, in combination with other research, to assess the potential for a functioning hydroelectric installation market to exist in Oregon.
Clackamas County, Oregon was selected as the study area. Located in the foothills of the Cascade Range, the county features substantial elevation changes, large numbers of streams, a combination of urban and rural land and water uses, and proximity to urban engineering and hydropower installation resources.
To make the study’s results relevant to the broadest possible audience, it was designed to:
• Identify potential hydro sites utilizing current property ownership boundaries.
• Screen stream resources for local, state, and federal environmental regulations.
• Generate site-specific power potential estimates, and
• Calculate approximate distances to interconnection infrastructure.
A GIS map of the county was constructed with layers for property ownership and boundaries, elevation, stream locations and federal and state protected areas. The layers included the following:
• Property ownership and boundary data available from Clackamas County.
• United States Geological Survey (USGS) 10m National Elevation Data (10m NED).
• Water courses (stream locations) available from the Pacific Northwest (Oregon & Washington) Hydrography Framework.
• Protected streams as determined by the Northwest Power and Conservation Council (a local regulatory agency).
• Electricity distribution infrastructure data provided by the local utility, Portland General Electric.
Utilizing these map layers, an initial analysis layer was created by filtering for privately owned lands outside of any protected areas (Figure 1).
Within the initial analysis layer watercourses crossing private lands were screened for environmental protections. Unprotected streams were added to the analysis area (Figure 2).
Using the 10m NED data, elevation contours were derived and overlaid on the map. Watercourses were intersected with the elevation contours and constant-elevation stream segments were created.
Contiguous stream segments with common ownership were joined and a length was attributed to the segments. To limit the number of analysis points, segments were eliminated if they did not have an elevation change of least 10m over a 350m horizontal run (Figures 3 and 4).
These steps yield point locations where unprotected watercourses exit privately owned properties. For each point, the stream segment’s elevation change across the property (head) has been calculated.
The resulting points were provided to the USGS to be run through the StreamStats application. StreamStats is a basin-specific GIS application capable of computing estimates of streamflow at ungauged sites based on basin delineation regression equations (Figure 5).
In some US states StreamStats is capable of estimating seasonal and/or average annual streamflow. The Oregon version of StreamStats is only capable of estimating peak streamflow events. The smallest maximum instantaneous streamflow event the Oregon version is able to estimate is a two-year peak flow event.
Peak flow events are generally not a good indicator of average annual or seasonal streamflow, but they are better than no data at all. In this case, the geographic, climatic, and stream monitoring characteristics of the study area allow some generalizations to be made from the StreamStats two-year maximum instantaneous flow data.
There are 10 streamflow gauging stations on unregulated rivers and streams within the Clackamas County study area. Data from these gauging stations was analyzed to determine the range of variation between average annual and peak streamflow events. Depending on the relative frequency of the peak event and the area of drainage involved, streamflow during peak events ranges from 12 – 129 times greater than average annual flow (see Table 1).
In the interest of creating conservative estimates, the upper end of this range was used to process the average annual streamflow at the identified study points. USGS estimated two-year peak flow values at the analysis points were divided by 129 in order to derive an average annual flow estimate. That flow estimate was again divided by 4 to reflect the on-the-ground reality in Oregon that state environmental agencies are unlikely to allow a run-of-river hydro power site to divert more than 25% of streamflows for hydro power production.
Using the estimated elevation change and the estimated average annual streamflow data, theoretical power was calculated for each point using the following equation:
P = power in kilowatts,
Q = flow in cubic feet per second,
H = head in feet,
11.81 = the constant for converting flow and head to kilowatts,
e = the efficiency of the plant, considering head loss in the pipeline and the efficiency of the turbine and generator, expressed by a decimal
For the purposes of conservative estimates H was multiplied by 0.8 to account for probable head losses and e was assumed to be 60%.
These estimates were used to create a visual representation of the potential capacity of the identified sites (Figure 6).
Following the power generation estimation, the individual sites were screened to determine their distance from power distribution lines. Energy Trust was only interested in sites that could potentially be integrated into the existing power grid. The local electric utility, Portland General Electric, provided data on power line locations and the phases available on those lines (single, two phase, or three phase). All sites with an estimated power potential greater than or equal to 1kW located within 2000ft (610m) of any power line, regardless of phase, were included in the results (Figure 7).
Following this process, the results were analyzed and checked for errors.
A total of 1812 potential site points were identified by the GIS analysis. Of those, 396 sites had a power potential greater than 1kW and were within 2000ft (610m) of a power line. Those sites were scrutinized to check for errors.
The error checking process identified some major problems. As an initial check, the identified points were overlaid with the protected streams layer to make sure sites were not incorrectly identified. In that test, many of the identified points appeared to be very close to protected streams (Figure 8).
A close up view shows that the identified sites are not actually on the protected streams, just nearby (Figure 9).
More closely checking these sites near to protected streams uncovered a major error in the streamflow estimation process.
The StreamStats application works by projecting peak flow events as they relate to precipitation and the area of drainage in a particular basin. StreamStats returns data indicating the number of square miles drained at the point of streamflow estimation. As shown above, many of the sites identified in this study were on small tributaries very close to a major stream or river. In many cases, StreamStats automated basin delineation process selected the major stream or river instead of the small tributary. Examining the drainage data, it is relatively easy to see points on small tributaries that do not correlate with large areas of drainage. The streamflow estimates at sites like these are for the main river and do not reflect the tributaries’ flow. This invalidates the results at sites where this occurred.
To determine the scope of impact of the above error, all 396 sites were examined to see if the reported area of drainage had a reasonable correlation with the apparent length of the tributary or stream. Of those 396 sites, 231 appeared to be inaccurately delineated by StreamStats. Those sites were eliminated from the study results (Figure 10).
A second error check was performed to determine if site ownership would preclude potential development. A total of 45 sites were eliminated due to poor ownership status. A majority of these sites were owned by the federal government and should have been eliminated earlier in the GIS process. A smaller portion of the sites were removed because they were owned by large timber companies. It was judged that these owners would be unlikely candidates for the development of a micro hydro project.
Of the remaining 120 sites, the largest was estimated to be 23kW. The next four largest sites were estimated at 9, 6, 5, and 4kW respectively. Eight 3kW sites and thirteen 2kW sites were also identified. The remaining 94 sites were all estimated to be 1kW.
A final check was performed on the sites greater than 4kW. None were within 2000ft (610m) of a three-phase power line, which may reduce development potential due to three-phase to single phase conversion inefficiencies (Figure 11).
Environmental protections across Oregon’s stream systems appear to acutely limit the potential to develop run-of-river micro hydro resources in the study area. These highly regulated stream systems are the legacy of the environmental, cultural, and recreational issues created by the Pacific Northwest’s large hydro installations.
This study found only four potential sites in Clackamas County with estimated power capacities greater than or equal to 4kW. However, study errors markedly limited accurate site identification and additional sites might have been found in the absence of these errors. Without further investigation it is impossible to know if other potentially viable sites exist in the county, but given the scale of regulation on the stream system any additional sites are likely to be few in number and small in power potential.
For Energy Trust, these results will focus the organization’s hydro outreach activities on potential sites where water is already allocated in an existing water right.
This GIS methodology appears sound with regard to estimating head at potential sites, but was prone to errors in estimating streamflow. Further iterations on this technique might improve performance and/or reliability.
Four factors could be tested to attempt to improve the study methodology:
• This study arbitrarily limited the analysis to streamflow segments that had an elevation change of at least 10m over a 350m horizontal run. The reason for this was to limit the overall number of points in this data-intensive analysis. This artificial limit may have eliminated some potential low-head sites with adequate flow while favoring higher-head sites with inadequate flow.
• Many potential sites identified in this process were located on tiny tributaries (less than 500m long). Estimates of power potential at these types of sites were often less than 1kW. Utilizing a lower resolution stream data set that focused on larger streams, not tiny tributaries, could have eliminated these low-power or no-power sites. Another method to reduce the identification of these types of sites could be to set a minimum stream length for inclusion in a study.
• Manually delineating stream basins in StreamStats could potentially eliminate major errors in the streamflow estimation process. Lack of time and resources prevented a test of this methodology. Utilizing a different stream flow estimation technique could also provide different results.
• Though not included in the current study, on-the-ground sampling could also be done to ground truth power potential estimates at sites that appear promising.
Variants on this methodology might be useful in other locations where StreamStats data exists or, utilizing a different streamflow estimation technique, in other nations to help identify potential micro hydro sites.
Jed Jorgensen is a Renewable Energy Project Manger at the Energy Trust of Oregon, Inc., an independent nonprofit organization dedicated to energy efficiency and renewable energy development. Email: [email protected]