Purpose and Scope

This report presents the methods and results of a study that directly monitored the streambed elevation at three selected scour critical bridge sites in Idaho (fig. 1). This report also discusses the project goals and provides a description of the three selected sites. Methods used to install, operate, and disseminate the sonar data are described. Site-specific geomorphic survey data are summarized, and we explain how those data were used to estimate scour using empirically derived pier scour equations. Results are presented for geomorphic and scour estimates; the real-time data served to the USGS National Water Information System (NWIS) database (NWIS; U.S. Geological Survey, 2024a, 2024b, 2024c), and the USGS ScienceBase data release (Fosness and Dudunake, 2024). A summary comparing the observed scour depths to the most recent pier scour estimates during peak flows at each site for each of the 3 years of the study is described, including factors which may cause variability between observed scour and pier scour estimates. Finally, considerations for future real-time monitoring and research are presented.

Site Selection and Summary

The three sites were selected based on the following three factors: (1) The site is currently listed by ITD as scour critical, (2) the site is co-located with an existing USGS streamflow gaging station, and (3) the site has diverse hydraulic and geomorphic conditions (fig. 1). The selected bridges provided a combination of varying characteristics including channel substrate, pier style, and flow regulation. All references to elevation have been updated, where necessary, to the North American Vertical Datum of 1988 (NAVD 88) for consistency.

Site Description of ITD Bridge Structure Number 19935—Payette River Near Letha, Idaho (USGS Station Number 13250000)

Idaho Transportation Department bridge structure number 19935, constructed in 1960, carries Idaho Boulevard (County Road 3830) over the Payette River and is located 1.0 mile (mi) east of Letha, Idaho (fig. 1; table 1). The bridge is co-located with USGS streamflow gaging station Payette River near Letha, Idaho (USGS station number 13250000). Idaho Transportation Department bridge structure number 19935 is currently listed as scour critical with a National Bridge Inventory Code (item 113) value of 3 (Federal Highway Administration, 2019).

The bridge has 9 spans (360 feet [ft] total length) with 2 steel shell pile abutments and 8 bents. Each bent includes 5 cylindrical (1.3-ft diameter) steel shell pile piers with concrete caps. The piers selected for this study were located at bent two (second pier from south abutment). Bent two had the highest observed depths and velocities based on direct observation by USGS. At the time of the initial site inspection, the angle the flow velocity intersects the piers (angle of attack) during low flow was nearly 90 degrees to the piers. Because of the high angle of attack, the sonars at bent two would be protected from debris by bents three through eight. Because the angle of attack varies with stream stage, and because bent two is in the thalweg at low to medium flows, bent two was still prone to collecting debris on the upstream face, although less than bents three to eight.

Two sonars were installed on January 15, 2020 at the upstream and downstream piers at bent two (fig. 2). Two sonars were installed because of the high angle of attack and the possibility of signal interference due to debris. The upstream sonar elevation was 2,283.27 ft and the downstream sonar elevation was 2,283.26 ft. The streambed elevation at the upstream and downstream sonar at the time of installation was 2,279.90 ft and 2,278.00 ft., respectively. The streambed elevation at the upstream side of bent two was 2,279.56 ft. The bridge design specified that the “steel piling shall have a penetration of not less than 15 ft below the original ground surface,” resulting in a minimum estimated pile tip elevation of about 2,264.56 ft (Fosness and Dudunake, 2024). The as-built pile tip elevation is unknown. Five test holes were cored and sampled at the bridge site and included in the design plans. Test hole number two was located at bent two. From test hole number 2: a layer of large gravel extended from the design surface (2,279.90 ft) down about 4.5 ft, a sand layer extended another 3.25 ft, and a sandy clay layer extended another 4.5 ft. The remaining test hole was sand.

Table 1.    

Selected real-time pier scour monitoring sites in Idaho, water years 2020–22 (U.S. Geological Survey, 2024a, 2024b, 2024c)

[Latitude and longitude referenced to North American Datum of 1983. Gage datum elevation referenced to North American Vertical Datum of 1988. Dates shown as month/day/year. USGS, U.S. Geological Survey; ft, foot]

Table 1.    Selected real-time pier scour monitoring sites in Idaho, water years 2020–22 (U.S. Geological Survey, 2024a, 2024b, 2024c)

Highway	USGS station number	Stream name	Latitude (decimal degrees)	Longitude (decimal degrees)	Gage datum elevation (ft)	Gage datum uncertainty (ft)	Sonar installation date
Bridge structure number 19935: Payette River near Letha, Idaho (USGS site number 13250000)
County Road 3830; Idaho Boulevard	13250000	Payette River	43.89664465	−116.6277345	2,276.158	0.029	1/15/2020*
Bridge structure number 27415: Boise River near Parma, Idaho (USGS site number 13213000)
Hexon Road	13213000	Boise River	43.78163599	−116.9723414	2,199.493	0.032	2/24/2020
Bridge structure number 30730: Saint Joe River at Calder, Idaho (USGS site number 12414500)
Elk Prairie Road	12414500	Saint Joe River	47.27419178	−116.1885463	2,175.514	0.089	1/30/2020

^*

Two sonar sensors installed.

The bridge design maximum water-surface elevation is 2,288.63 ft. This elevation correlates to a current gage height of 12.47 ft (local USGS streamflow gaging station datum) and a streamflow of about 7,230 cubic feet per second (ft³/s; U.S. Geological Survey, 2024a).

The Payette River is regulated by several upstream dams including Black Canyon Diversion Dam, Deadwood Dam, Cascade Dam, and Lardo Dam. As a result, traditional peak streamflow statistical methods do not provide reliable peak flood frequency flow estimates. The Bureau of Reclamation provided regulated flood frequency statistics for the reach below Black Canyon Diversion Dam, located 13.4 mi upstream of the bridge. The 1-percent annual exceedance probability (AEP; commonly referred to as the 100-year recurrence interval) regulated flow estimate is about 22,800 ft³/s, and the 0.2-percent AEP (commonly referred to as the 500-year recurrence interval) flow estimate is about 43,100 ft³/s (Jonathan Rocha, Bureau of Reclamation, written commun., 2023). Overbank flow occurs at a gage height of about 17.85 ft (local USGS streamflow gaging station datum; elevation 2,294 ft). The corresponding flow is estimated at 29,600 ft³/s. Low chord (lowest elevation of the bridge superstructure) is at an elevation of 2,294.71 ft; pressure flow (submerged low chord) at the bridge is unlikely because overbank flow will exceed the elevation of the road to the south, flanking the bridge. The flow at a water surface elevation immediately below the low chord elevation on the bridge (2,294.71 ft) is about 33,900 ft³/s (U.S. Geological Survey, 2024a).

The streambed at the bridge is composed of sand, gravel, and some silt, and is generally free of vegetation. The main channel is straight near the bridge. As noted in the USGS streamflow gaging station description, the downstream bed elevation gradually changes, even at base flow, and requires regular stage-discharge relationship shifts (Fosness and Dudunake, 2024).

Site Description of ITD Bridge Structure Number 27415—Boise River Near Parma, Idaho (USGS Station Number 13213000)

Idaho Transportation Department bridge structure number 27415, constructed in 1954, carries Hexon Road over the Boise River and is located 0.4 mi south and 1.2 mi west of Parma, Idaho (fig. 1; table 1). The bridge is co-located with USGS streamflow gaging station Boise River near Parma, Idaho (USGS station number 13213000). Idaho Transportation Department bridge structure number 27415 is currently listed as scour critical with a National Bridge Inventory Code (item 113) value of 3 (Federal Highway Administration, 2019).

The bridge length is 331 ft and has 11 spans with 10 bents. Each bent includes 6 timber piles enclosed by wooden planks. In November 2017, bent 2 was rehabilitated with piles 1 (downstream) and 6 (upstream) replaced with hollow structural pipe (12.75-inch wide by 0.5-inch thick). Additionally, bents 1, 3, 4, and 5 had 10-gauge galvanized corrugated metal pipe nose armor installed on the upstream side. Bents 1–4 (from the west abutment) are in the main Boise River channel with bents 5–9 (from the west abutment) in a secondary channel inundated only during higher flows. The piers selected for this study were located at bent two (from the west abutment). Personnel from the USGS surveyed the channel along the upstream face of the bridge prior to installing the sonar. The sonar soundings and surveys indicated that the streambed was uneven near bent two (the location of the sonar).

The sonar was installed on February 24, 2020, on the upstream side of bent two at an elevation of 2,205.41 ft (fig. 3). The streambed elevation at the time of sonar install was about 2,201.8 ft. The streambed elevation near bent two from the design notes was about 2,206.76 ft., which is about 5 ft higher than the current streambed elevation. The bridge design noted, “all [timber] piling shall be driven to a minimum penetration of 15 feet below stream bed elevation,” which results in a designed highest pile tip elevation of 2,191.76 ft (Fosness and Dudunake, 2024). The as-built pile tip elevation is unknown, and no test hole data were available noting the particle size by depth.

The bridge design maximum water-surface elevation is noted in the design documents as 2,215.76 ft., which is the high-water elevation from a flood that occurred in 1943. Streamflow in the lower Boise River is regulated by Lucky Peak Dam 60 mi upstream of the bridge, along with numerous irrigation diversions and returns. Regulated peak streamflow statistics were provided by the U.S. Army Corps of Engineers for the 10-percent, 1-percent, and 0.2-percent AEPs. The regulated peak streamflow statistics were calculated for the USGS streamflow gaging station Boise River at Glenwood Bridge near Boise, Idaho (USGS station number 13206000; 43.7 mi upstream of the bridge), but the peak flow is expected to be about the same with only minor gains and losses during peak flows. Regulated peak streamflow statistics are 7,500 ft³/s for the 10-percent AEP, 16,600 ft³/s for the 1-percent AEP, and 34,800 ft³/s for the 0.2-percent AEP (U.S. Army Corps of Engineers, 2015). Streamflow less than 3,800 ft³/s remains within the main (low flow) channel, while bank-to-bank flows occur at a streamflow exceeding 7,100 ft³/s.

Low chord on the bridge is at an elevation of 2,216.95 ft; pressure flow could occur at any water-surface elevation above the low chord elevation due to the main channel containing the flow. The maximum computed flow at the USGS streamflow gaging station (from the stage-discharge rating curve) is 11,500 ft³/s occurring at a water surface elevation of 2,215.51 ft. The maximum computed flow with rating extension to the low chord bridge elevation is 14,900 ft³/s. Channel banks are vegetated rip-rap dikes with varied amounts of growth and contain all streamflow up to 11,500 ft³/s (the maximum discharge estimated at the USGS streamflow gaging station). Both the 1-percent AEP (38,400 ft³/s) and the 0.2-percent AEP (16,600 ft³/s) would likely cause pressure flow conditions at the bridge. The channel bed is composed of sand, gravel, and cobble and is generally straight for several hundred feet upstream and downstream from the bridge (gage).

Site Description of ITD Bridge Structure Number 30730—Saint Joe River at Calder, Idaho (USGS Station Number 12414500)

Idaho Transportation Department bridge structure number 30730, constructed in 1990, carries Elk Prairie Road over the Saint Joe River and is located 0.1 mi south of Calder, Idaho (fig 1; table 1). The bridge is co-located with USGS streamflow gaging station Saint Joe River at Calder, Idaho (USGS station number 12414500). Idaho Transportation Department bridge structure number 30730 is currently listed as scour critical with a National Bridge Inventory Code (item 113) value of 3 (Federal Highway Administration, 2019).

The bridge has 3 spans (296 ft total length) with 2 steel pile abutments and 2 bents. Each bent includes 10 steel shell pile piers with concrete caps measuring 29.2 ft long by 3.5 ft wide. The pier selected for this study was located at bent two (northernmost bent). The sonar was installed on January 30, 2020, at an elevation of 2,180.9 ft. (fig. 4). The streambed elevation at the time of installation was 2,177.7 ft. The streambed elevation at bent two, as noted in the design notes, was about 2,177.6 ft with the concrete extending below the ground surface to an elevation of 2,173.35 ft. Rip-rap, an angular rock erosion countermeasure (unknown size) was added around the bents (5 ft along the sides and extending 10 ft upstream and downstream). The bridge design noted that pier 2 minimum pile tip elevation shall be driven to 2,097.56 ft. The as-built pile tip elevation is unknown. Four test holes were cored and sampled at the bridge site and included in the design plans. Test drill hole number one was located about 37 ft south of bent two and described as “Silty, sand and gravel with cobbles” to a bed elevation of 2,160.56 ft (Fosness and Dudunake, 2024). The remaining test hole was sand.

The bridge design maximum water-surface elevation (for the designed 1-percent AEP) was 2,191.50 ft, almost 2 ft below the bridge low chord elevation of 2,193.6 ft. Flow upstream of the bridge site is unregulated, and peak flow was estimated for the 10-percent, 1-percent, and 0.2-percent AEP using the USGS StreamStats program (U.S. Geological Survey, 2019). The 10-percent AEP flow estimate is 26,100 ft³/s; 1-percent AEP flow estimate is 39,900 ft³/s; and the 0.2-percent AEP flow estimate is 49,700 ft³/s. The 1-percent AEP design elevation correlates to a streamflow of about 40,000 ft³/s. Overbank flow, as noted by the USGS station description, occurs at about 14,100 ft³/s. Overbank flow combined with the bridge being elevated above the surrounding floodplain results in minor contraction of the flow under the bridge. Additionally, the overbank flow prevents pressure flow at the bridge (low chord elevation is 2,193.6 ft). The main channel surface substrate is gravel and cobble, and the channel is generally straight in the vicinity of the bridge and streamflow gaging station.

Real-Time Pier Scour Monitoring

Real-time (15-minute interval) streambed elevation monitoring provided a bridge scour countermeasure by remotely observing bridge scour conditions and recording the streambed elevation data. Directly monitoring streambed elevation is the most accurate method to determine if a bridge is approaching a scour critical depth (Ayres Associates, 2004a). Real-time instrumentation and data collection methods used for this study paralleled those described at length in Henneberg (2018). Airmar EchoRange Single Frequency Smart Transducers (sonars; Airmar Technology Corporation, 2019), were mounted within enclosed and custom fabricated steel housings and directly mounted to the piers. The sonars were mounted on the piers so that they would measure the lowest streambed elevation at the nose of the piers. The elevation of the measuring point on each sonar was derived using survey leveling from the sonar measurement points to USGS streamflow gaging station reference marks (Kenney, 2010). A global navigation satellite system (GNSS) survey determined the elevation of each reference mark referenced to the NAVD 88. The USGS reference mark was surveyed following the techniques and methods in Rydlund and Densmore (2012). The elevation of the sonar measuring point (zero depth) was computed by subtracting the vertical distance from the reference mark to the measuring point on the sonar. The sonar power and data transmission lines were run back to the USGS streamflow gaging station along utility corridors or the outside of bridge railings. At the gaging station shelter the National Marine Electronics Association 0183 (NMEA 0183) interface standard data string was converted to serial digital interface at 1,200-baud (SDI–12) protocol data format using a NMEA 0183 to SDI-12 protocol converter (Henneberg, 2018). The protocol converter was connected to the data collection platform (DCP) using standard SDI-12 wiring conventions. The protocol converter was queried by the DCP for raw depth below transducer values using standard SDI-12 commands. Every 15 minutes after the hour the DCP queried the protocol converter for the instantaneous depth below transducer value from the sonar and computed the streambed elevation by subtracting the depth below transducer from the surveyed elevation of the sonar. The 15-minute streambed elevation values were logged to the DCP and transmitted hourly via Geostationary Operational Environmental Satellite (GOES) network to the National Water Information System (NWIS) data repository (https://doi.org/10.5066/F7P55KJN; Henneberg, 2018). The telemetered 15-minute streambed elevation was then published to U.S. Geological Survey National Water Information System (U.S. Geological Survey, 2024a, 2024b, 2024c).

Geomorphic Site Survey and Hydraulic Assessment

Geomorphic site surveys provided site specific parameters required for the hydraulic assessment. The real-time bed elevation provided measurable observations, but the hydraulic assessment was a useful method to compare to the (1) observed scour data and (2) selected flood frequency flows. This section describes the at-site geomorphic data collection along with the methods used to develop the hydraulic assessment.

Real-time Streambed Elevation Monitoring Data Quality Assurance

Quality assurance methods were established prior to the study to ensure consistent and accurate techniques and methods were applied to the real-time streambed elevation monitoring at each of the study sites. Gage datum surveys established the vertical reference to NAVD 88 at each site and the uncertainty associated with the GNSS derived elevation was documented (Rydlund and Densmore, 2012). Elevations at the sonars were established using standard leveling techniques from the gage reference marks (with NAVD 88 elevation; Kenney, 2010). Computed streambed elevations were periodically verified on site by USGS hydrographers by sounding from a known reference point. Sounding quality assurance methods included comparisons with measured depths from an acoustic Doppler current profiler and physical depth measurements (using a Columbus weight and reel, weighted tape, or survey rod; fig. 5).

Data gaps existed within the real-time data record largely due to a combination of debris, ice, power malfunctions, and the sonars coming out of the water at low flows. Debris accumulation generally occurred as large woody debris entangled on the pier and blocked the sonar signal, floating vegetation caught on the pier or sonar, or a combination of each. Every effort was made to remove debris when it was observed during routine site visits or when viewed in the real-time data. Occasionally, large woody debris could not be removed from the pier. Ice accumulation was not as prolific as debris but did occur occasionally. Unexpected loss of power occurred occasionally and resulted in a loss of data during the outage. Power was restored at the gaging station as soon as possible. The sonars were installed as low as reasonably possible, although they did come out of the water occasionally at low flows.

In an effort to filter out erroneous data collected from the sonar data, minimum and maximum elevation values, or thresholds, were established at each site. In general, high thresholds were used to filter data affected by debris (large woody and floating vegetation) and periods when the sonar was out of water or affected by ice; low thresholds were used to filter reflected sonar data resulting in extreme low values.

The sonars used for this study applied the default and constant speed of sound underwater (4,921 ft per second; Airmar Technology Corporation, 2019) to measure the depth. The speed of sound underwater is known to change based on water temperature, possibly resulting in erroneous depth measurements (Xylem, 2018). Water temperature was measured and recorded from the sonar and transmitted with the streambed elevation data. The effect of water temperature on depth was analyzed at each site to assess the need for a water temperature correction. We determined that no corrections were needed to the sonar measurements due to (1) shallow depths and (2) low variability in water temperatures. These factors resulted in a depth uncertainty of less than 0.1 ft., within the accepted accuracy of the sonar (Airmar Technology Corporation, 2019). The quality-assurance depth checks also confirmed water temperature corrections were not necessary for the study sites.

Real-Time Streambed Elevation and Scour Observations for ITD Bridge Structure Number 19935—Payette River Near Letha, Idaho (USGS Station Number 13250000)

Real-time streambed elevation was monitored at the upstream and downstream portion of bent two. Real-time streambed elevation and streamflow data are available at NWIS (U.S. Geological Survey, 2024a) for the study duration. Data gaps in the streambed elevation values at the Payette River near Letha, Idaho occurred due to miscellaneous gage malfunctions, erroneously high and low values, and the sonar coming out of the water during low flow periods. NWIS defines a data gap any time values are missing for more than 2 hours. Many erroneously high values attributed to aeration in the beam path of the sonar were screened and caused data gaps in the record. Periodically, reflected sonar data caused erroneously low values, which were also screened from the record. Data gaps due to aeration or reflected sonar data do not exceed 5 hours. Quality assurance depth checks collected periodically at the site provide an uncertainty estimate for the observed scour data. The observed pier scour uncertainty for the upstream and downstream piers ranged from 0.1 to 0.2 ft (table 3). Daily mean streamflow during the study period (water years 2020–22) was generally at or below the median daily statistic for the site. Instantaneous peak streamflow ranged from a low of 4,600 ft³/s (April 5, 2021) in water year 2021 (ranked 36 within the 38-year peak flow record) to a high of 12,800 ft³/s (June 13, 2022) in water year 2022 (ranked 18 within the 38-year peak flow record). The water year 2022 peak flow was larger than the average and median peak flows at the site of 12,200 ft³/s and 11,900 ft³/s, respectively. In general, the daily mean and instantaneous peak flows were below or near average for the duration of the study period (U.S. Geological Survey, 2024a; fig. 6).

Streambed elevation data collected during the water year 2020 peak runoff period (March–June) did not indicate significant scour at the upstream or downstream sonar (fig. 6). Observed peak-to-peak scour for water year 2020 at the upstream and downstream sonars were 1.5 ft and 0.9 ft (±0.1 ft), respectively. Water year 2021 flow was well below average resulting in an aggrading (increasing) streambed elevation within the March–June runoff period. The channel bed elevation aggraded from March to June resulting in a negative peak-to-peak scour value (fill) at the upstream and downstream sonar (−0.6 ft and −0.7 ft [±0.2 ft], respectively). While the peak flows were largest during the runoff period in water year 2022, the resulting observed peak-to-peak scour depths were similar to water year 2020 at 1.4 ft (±0.2 ft; upstream) and 0.8 ft (±0.2 ft; downstream) (fig. 5; table 3). The minimum pile tip elevations given in the design plans suggest that the scour depth to the pile tips would need to be about 16 ft for the upstream pier and about 14 ft for the downstream pier. The subsurface particle gradation (from test hole number 2 in the bridge design), while not measured as part of this study, suggests a coarse bed to an elevation of 2,274.9 ft, or a depth of about 5.1 ft (upstream) and 3.3 ft (downstream).

Coarse Bed equation calculated scour values generally overestimated scour when compared to the observed streambed elevation data. Pier scour was calculated at a single location at the upstream side of the bridge, and the same pier scour value was used to compare against the upstream and downstream observed scour values. For water year 2020 the calculated scour overestimated the scour depth by 2.1 ft at the upstream sonar and 2.7 ft at the downstream sonar. Calculated scour for water year 2021 was 1.6 ft and overestimated the observed scour by 2.2 ft at the upstream pile and 2.3 ft at the downstream pile. The largest calculated scour occurred in water year 2022 at 4.5 ft; again, overestimating scour by 3.1 ft (upstream) and 3.7 ft (downstream). In general, the calculated scour values produced comparable scour estimates to the observed peak-to-peak streambed elevation change (table 3). Both the observed streambed elevation changes and calculated scour values were less than the scour depth to the minimum pile tip elevation at the upstream and downstream piers.

Real-Time Streambed Elevation and Scour Observations for ITD Bridge Structure Number 27415—Boise River Near Parma, Idaho (USGS Station Number 13213000)

Real-time streambed elevation was monitored at the upstream portion of bent two. Real-time streambed elevation and streamflow data are available at NWIS (U.S. Geological Survey, 2024b) for the study duration. There are many data gaps in the streambed elevation data at the Boise River near Parma due to heavy aquatic vegetation growth in the channel and drifting aquatic vegetation that collected on the pier and sonar. Aquatic vegetation was cleared during site visits; however, data gaps generally increased throughout the growing season. The observed pier scour uncertainty varied by up to 0.5 ft due to an uneven streambed at the pier (table 3). Soundings and a channel cross section survey prior to sonar installation indicated that the streambed is 1.1 to 2.3 ft higher immediately upstream of the bridge than directly underneath the sonar. Prior to sonar installation, soundings were made on both sides of the pier. The streambed elevation on the left bank side of the pier was 2,203.0 ft and the streambed elevation on the right bank side of the pier was 2,201.8 ft. Soundings on March 16, 2020, also revealed differences from the left to right bank sides of the pier. Due to the highly turbid nature of the river, it is unclear if the difference in the soundings from either side of the pier were due to debris on the streambed, erroneous soundings on the bridge structure, or an actual difference in the streambed elevation. Approach flow angle of attack does not suggest there would be a scour hole on one side of the pier or the other (table 2). Daily mean streamflow during the study period was generally at or below the median daily statistic for the site. Annual instantaneous peak streamflow ranged from a low of 2,090 ft³/s (May 20, 2020; ranked 34 within the 47-year peak flow record) to a high of 3,270 ft³/s (June 13, 2022; ranked 29 within the 47-year peak flow record). The average and median peak flows of 4,960 ft³/s and 5,780 ft³/s, respectively, are larger than each of the peak flows during the study period. In general, both the daily mean and instantaneous peak flows were below or near average for the duration of the study period (U.S. Geological Survey, 2024b).

Streambed elevation data for water year 2020 showed a scour pattern in late May through mid-June resulting in a peak-to-peak observed scour of 1.8 ft (±0.5 ft; fig. 7). water year 2021 resulted in the channel aggrading until the peak flow, scouring through mid-June, then rapidly filling through late June. While there was some observed scour during runoff in 2021, the net observed peak-to-peak change in streambed elevation resulted in 2.2 ft (±0.2 ft) of fill. Water year 2022 experienced the largest peak flow, and the streambed elevation generally remained stable until the peak flow in June resulting in 1.3 ft (±0.5 ft) of peak-to-peak scour (table 3). Throughout the study period the streambed elevation at bent 2 aggraded. The average streambed elevations in 2020, 2021, and 2022 are 2,202.5, 2,202.9, and 2,203.3 ft, respectively. The minimum design pile tip depth given in the design plans suggest a scour depth to the minimum pile tips of about 11 ft (table 3).

Coarse Bed-calculated scour values generally increased for each year following the same trend as the peak flow. The calculated scour for water year 2020 was 1.4 ft.; underestimating the observed scour by 0.4 ft. Similarly, water year 2021 produced a calculated scour estimate of 1.5 ft; overestimating by 3.7 ft compared to the observed scour. Likely because of the increased peak flow in 2022, the calculated scour in water year 2022 was 1.6 ft, an overestimate of 0.3 ft compared to the observed scour (table 3). While the water year 2021 scour results overestimated scour by over 3 ft due to aggradation at the pier that year, the calculated scour estimates for water years 2020 and 2022 were within 0.5 ft. of the observed scour. Both the observed and calculated scour values were less than the scour depth to the minimum pile tip elevation.

Real-Time Streambed Elevation and Scour Observations for ITD Bridge Structure Number 30730—Saint Joe River at Calder, Idaho (USGS Station Number 12414500)

Real-time streambed elevation was monitored at the upstream portion of bent number 2. Real-time streambed elevation and streamflow data are available at NWIS (U.S. Geological Survey, 2024c) for the study duration. There are periodic data gaps of less than 24 hours throughout the streambed elevation record at the Saint Joe River at Calder, Idaho due to erroneously high and low values caused by reflected sonar data off the angular rip rap armoring the pier. Longer duration data gaps, up to 46 days, during high flow and late spring, are caused by the filtering of erroneously high values caused by debris that collected on the nose of the pier. Special visits were made to clear debris during peak flows; however, debris-affected values were not urgent during the flow recession so special visits were not prioritized. Small debris during the flow recession would generally clear the nose of the pier during the next peak flow event. There are periodic data gaps in the late summer due to the sonar coming out of the water and during the winter due to ice on the sonar. Occasional equipment malfunctions also caused short duration data gaps. Daily mean streamflow during the study period was generally at or below the median daily statistic for the site. Annual instantaneous peak streamflow ranged from a low of 9,210 ft³/s (May 2, 2021; ranked 93 within the 105-year peak flow record) to a high of 17,000 ft³/s (May 7, 2022; ranked 40 within the 93-year peak flow record). Only the peak flow in water year 2022 exceeded the average and median peak flows of 16,600 ft³/s and 15,400 ft³/s, respectively. Except for water year 2022, the mean daily and peak flows were generally below average (U.S. Geological Survey, 2024b).

Water year 2020 produced an observed peak-to-peak scour of 0.9 ft (±0.5 ft; fig. 8). Water year 2021 resulted in the channel generally aggrading until the peak flow in May. Overall, the channel largely aggraded, with a peak-to-peak scour of 0.8 ft (±0.2 ft). Water year 2022 produced an average peak flow of 17,000 ft³/s but the streambed elevation experienced only minor changes with a peak-to-peak scour result of 1.0 ft (±0.2 ft; table 3). The observed pier scour uncertainty varied from 0.2 to 0.5 ft; this was likely because the streambed elevation at the pier was uneven due to a scour hole on the upstream right bank side of the pier and large angular rip-rap at the pier. Quality-assurance soundings collected directly upstream of the pier nose were less than soundings directly at the sonar and may have resulted in increased uncertainties due to measuring away from the sonar sounding location. Additionally, erroneous sonar soundings were attributed to beam scattering caused by the angular rip-rap at the pier and an uneven streambed surface. The streambed elevation data obtained from this study (about 2,178 ft) suggests that a scour depth of about 80 ft would be required to reach the minimum pile tip elevation (2,097.56 ft).

Coarse Bed calculated scour values ranged from 1.7 ft (water year 2021) to 2.9 ft (water year 2022; table 3). The calculated values for water years 2020 and 2021 were within about 1.4 foot of the observed scour. Water year 2022 calculated pier scour overestimated by 1.9 ft compared to the observed scour. The water year 2022 peak flow (17,000 ft³/s) was above average for the site.

Risk and Accuracy Assessment

HEC-18 design pier scour equations were not developed to predict scour depth but rather predict the maximum probable scour depth for design applications. While conservative scour estimates are desirable for safety reasons, overly conservative scour estimates may lead to (1) over-designed bridge foundations and (2) overestimation of predicted scour at existing bridges. For each case, overly conservative scour estimates can lead to an increase in design, construction, and operating cost that is not economically acceptable. The observed scour data from this study were compared to two HEC-18 design equations (Coarse Bed and HEC-18 general pier scour; fig. 9) and summarized in terms of risk (reliability index) and accuracy (relative root mean square error).

The observed pier scour depth was compared to each depth calculated by the Coarse Bed and HEC-18 general pier scour equations (Federal Highway Administration, 2012a, 2012c, 2016). A comparison of the observations and calculated results for each HEC-18 design equation provided a (1) reliability index and (2) relative root mean square error for the dataset (table 4).

The reliability index (RI) describes the risk of underpredicting the pier scour depth (Federal Highway Administration, 2012a, 2012c, 2016). The Federal Highway Administration described a target reliability index equal to 2.0 based on, “an assessment of the reliability of current practice” (Federal Highway Administration, 2016, p. 22). An RI greater than 2.0 indicates a lower risk of underpredicting scour, whereas an RI less than 2.0 indicates a higher risk of underpredicting scour (Federal Highway Administration, 2012c, 2016). The RI was calculated using the same approach as FHWA to compare the RI for the Coarse Bed and HEC-18 general pier scour equations. The results from this study were an RI of 1.2 (higher risk of underpredicting scour) for the Coarse Bed design equation and 2.2 (lower risk of underpredicting scour) for the HEC-18 general scour design equation (table 4).

The relative root mean square error (RRMSE; table 4) describes the overall error associated with each method compared to the observation to evaluate scour design equation prediction performance. The Coarse Bed design equation overpredicted scour depths (RRMSE =2.2) as intended by the design equation. The Coarse Bed design equation generally overpredicted scour by about 2.5 times less than the HEC-18 general pier scour equation for data collected during the 3-year study period.

Considering the risk (RI) and the accuracy (RRMSE), the Coarse Bed design equation provided a more accurate result than the HEC-18 general pier scour equation but less reliable at overpredicting the scour depth compared to the general scour equation. The risk and accuracy assessments are based on 12 observations collected from this study. Additional field data collected over a large range of flow observations are needed to perform a more thorough assessment to compare the field observations with the predictions from the HEC-18 pier scour design equations (Coarse Bed and HEC-18 general).