Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
4 NCHRP LRD 82 projects, State DOTs cannot avoid liability for failure to perform routine maintenance of highway slopes. II. UNSTABLE SLOPE MANAGEMENT PROGRAMS A. History Unstable slope management programs generally relate back to the work of Canadian professional engineer Duncan C. Wyllie, who proposed methods to rationally allocate limited maintenance funds available for highway slope stabilization.14 As early as 1975, Wyllie proposed grouping rockfall sites into five categories ranging from category âAâ (those requiring im- mediate maintenance) to category âEâ (those not requiring maintenance), based on both the potential for a rockfall event and the expected cost of a rockfall event at each site.15 In 1980, Wyllie proposed a more detailed evaluation of âthe stability conditions for each slope on a numerical point rating from very high to very low probability that a rockfall will oc- cur,â based on rockfall history as well as âphysical and geological characteristics of the slopesâ including âthe length and spacing of the natural fractures in the rock, their strength characteris- tics and orientation with respect to the slope face, groundwater pressures,â etc.16 In addition to evaluating the likelihood that a slope failure would occur, Wyllie recommended considering the costs in the event of a slope failure, including âthe costs of a delay caused by a major rockfallâ such as âinterruption to traf- fic,â the direct cost of roadway repair including debris removal, the direct cost of property damage and personal injury or death, and also âindirect costs such as the lost wages of those injured, engineering studies of stability conditions, and legal fees in the event of a court case.â17 The âexpected costâ of failure for a given slope would be the product of its probability of failure and the aforementioned costs in the event of failure. Highway maintenance departments could then determine the most cost- effective course of action by comparing the cost of different mit- igation alternatives (e.g., ranging from no mitigation, to scaling loose rock from the slope face, to installing protective devices such as rock anchors, catchment ditches, and barrier walls) with the attendant reduction in the expected cost of a failure for a given slope.18 Although Wyllieâs 1980 proposal provided a framework for making decisions regarding maintenance or mitigation of haz- ardous slopes, it was limited because there was no way to ac- curately calculate either the probability of slope failure or the costs of a slope failure. In addition, because a catastrophic slope 14 Pierson, supra note 1, at 6. 15 C.O. Brawner & Duncan Wyllie, Rock Slope Stability on Railway Projects, Proceedings of the American Railway Engineering Association Regional Meeting, Vancouver, B.C. (1975). 16 Duncan C. Wyllie, N.R. McCammon & W. Bramund, Planning Slope Stabilization Programs by Using Decision Analysis, 749 Transp. Res. Rec.: J. Transp. Res.. Bd. 35 (1980). 17 Id. 18 Id. at 36â38. â¢ Personal injury or property damage suffered by travelers on the highway who impact the falling or sliding debris;10 â¢ Personal injury or property damage incurred by trav- elers due to the dangerous condition of an unrepaired highway after a slope failure event; â¢ Personal injury or property damage to others located downslope from the highway, where the slope failure is attributable to the State DOT (e.g., resulting from con- struction activities or failure of a mitigation measure); and â¢ Physical damage to the highway property resulting from the slope failure, requiring the State DOT to undergo the expense of repair or at least cleanup. Personal injury or property damage resulting from high- way slope failures often results in tort claims made against State DOTs.11 Although State DOTs have a duty to maintain highways in a reasonably safe condition for travel,12 it is gen- erally understood that it is not feasible to completely elimi- nate the risk of highway slope failure.13 Courts, administrative boards, or other tribunals considering such tort claims will al- most always undertake an analysis of governmental immunity. Section III provides an overview of tort immunity for State DOTs, with a focus on the discretionary function exception to most waivers of immunity for State DOTs. Section IV cov- ers reported court decisions applying these legal principles to highway slope failures. In an effort to manage the cost and risk (legal and other- wise) of highway slope failures, many State DOTs have adopted unstable slope management programs, such as the Rockfall Hazard Rating System (RHRS) developed with grant funding from the Federal Highway Administration (FHWA). SectionÂ II provides an overview of unstable slope management programs. SectionÂ V examines how unstable slope management programs influence the outcome of tort claims against State DOTs in cases involving highway slope failures, including how they influence the courtâs analysis of immunity and negligence. While unstable slope management programs can help State DOTs avoid liabil- ity for decisions to defer capital-intensive slope remediation 10 Thompson, supra note 3, at 8 (âThe traveling public and their vehicles rarely notice or come into contact with these features unless there is a critical condition state or failure.â); Beckstrand et al., supra note 6, at 8 (âRockfall reaching and/or coming to rest in the roadway affects users by introducing sudden and unexpected obstacles requiring quick reaction to avoid.â). 11 Beckstrand et al., supra note 6, at 8 (â[T]hrough the research teamâs experience with rockfall-related legal claims, falling rock enter- ing vehicles at highway speeds is a principal cause of rockfall-related fatalities.â). 12 R.P. Davis, Duty as Regards Barriers for Protection of Automobile Travel, 173 A.L.R. 626 (1948). 13 Woodard, supra note 6, at 6 (âWith the thousands of miles of potentially problematic roadways in the United States, it is an over- whelming task to properly design all rock slopes to ensure 100% safety for travelers. .Â .Â . [I]t is generally not feasible to have all slopes designed to obtain 100% safety due to problems such as the lack of adequate funding or a limited right of way.â).
NCHRP LRD 82 5 ranked based on a benefit-cost analysis of the proposed reme- dial construction project.25 Beginning in 1984, the Oregon DOT began planning development of the RHRS, modeled largely on Wyllieâs research, in order to better account for rockfall poten- tial, rather than just rockfall history, of all rockfall sites.26 For purposes of RHRS development, a rockfall site was considered to be âany uninterrupted slope along a highway where the level and occurring mode of rockfall are the same.â27 Two slope failure modes were considered: (1)Â where rockfall occurs along natural structural discontinuities (such as joints or bedding planes), and (2)Â where rockfall occurs as a result of differential erosion or âoversteepening,â which is common in excavated slopes (where the instability results from the exposure of an underlying, softer rock layer to weathering).28 In developing the RHRS, the Oregon DOT initially used a variation of Wyllieâs original 1975 scoring system to group haz- ardous slopes by preliminary ratings, from class âAâ (high risk) to class âCâ (low risk), based on a subjective evaluation.29 After performing the preliminary ratings, the Oregon DOT used a variation of Wyllieâs 1987 exponential scoring system to perform a detailed rating of all class A slopes, with this detailed rating becoming the initial âprototypeâ of the RHRS.30 Funding from an FHWA Highway Planning and Research (HPR) pooled-fund study grant in 1989 (which included HPR funds from a number of State DOTs) allowed the Oregon DOT to complete software development and test the RHRS at over 3000 sites.31 This âshake- down testâ32 of 3000 sites resulted in some additional modifica- tions and deviations from Wyllieâs 10-category scoring system, as the Oregon DOT added and removed some categories, and revised the scoring for other categories, in an attempt to better prioritize the rockfall sites.33 The RHRS ultimately ârate[d] each site in 12 categories, including slope height, rockfall history, cli- mate and presence of water, the amount of traffic and structural condition.â34 The RHRS rating is not presented as a rigorous calculation, as it does not purport to predict the likelihood of rockfall or the likely cost, damage, or injury due to a rockfall event. In- stead, the RHRS rating generally identifies the most hazardous slopes. It has long been recognized that, although the highest hazard slopes generally receive the highest RHRS ratings, âoften 25 Pierson, supra note 1, at 13. 26 Id. at 6. 27 Id. at 7. 28 Id. at 10. 29 Id. at 7; see also Lawrence A. Pierson & Robert Van Vickle, Rockfall Hazard Rating System Participantâs Manual, Rep. No. FHWA-SA-93-057, 3 (1993) [hereinafter RHRS Manual]. 30 Pierson, supra note 1, at 6. 31 Id. 32 Lawrence A. Pierson & Robert Van Vickle, Report on the Shakedown Test of Oregonâs Rockfall Hazard Rating System, Rep. No. FHWA-OR-EG-89-01 (1989). 33 Pierson, supra note 1, at 6. 34 Richard L. Hill, Landslide, Portland Oregonian, Oct. 19, 1995, at 808. failure is a relatively low probability event at any given site, this proposed approach resulted in relatively low âexpected costâ cal- culations for all slopes and thus did not effectively differentiate between slopes. Therefore, in 1987, Wyllie published a revised rating procedure for prioritizing rockfall sites.19 This procedure involved scoring each slope based on 10 factors or categories, some related to the potential for rockfall and some related to the cost of a rockfall event at the site.20 A highway maintenance department would not have to rigorously calculate the score in each category but instead would perform an often subjective evaluation of the slope to assign it a qualitative score in each category ranging from low to high severity.21 An exponential scoring system was used, such that a low severity score was as- signed a numeric value of 30, or 1; a medium-low severity score was assigned a numeric value of 31, or 3; a medium severity score was assigned a numeric value of 32, or 9; a medium-high severity score was assigned a numeric value of 33, or 27; and a high severity score was assigned a numeric value of 33, or 81.22 Once the slope was evaluated for each of the 10 categories, the individual quantitative scores for each category (ranging from 1 to 81) would be summed together, resulting in a total score for each slope ranging from 10 to 810. The slopes could then be compared and ranked based on total scores, with the high- est total scores generally representing the slopes most in need of attention. The appeal of Wyllieâs ranking procedure included the fact that although it did not require detailed engineering analysis or calculations for each slope, the exponential scoring system allowed for a meaningful quantitative differentiation be- tween slopes. 1. Oregon Department of Transportation Rockfall Hazard Rating System The publication of Wyllieâs rating system in 1987, combined with the increasing availability and usage of personal computers by State DOTs at the time, inspired the development of a num- ber of computerized unstable slope management programs.23 At the forefront of these was the RHRS developed by the Oregon DOT with FHWA funding.24 For many years, the Oregon DOT had maintained a prior- ity list for proposed rockfall construction projects, containing about 100 sites identified as having a rockfall history involving accidents and/or excessive maintenance costs, with the sites 19 Duncan Wyllie, Rock Slope Inventory System, Proceedings of the FHWA Rockfall Mitigation Seminar and 13th Northwest Geotechnical Workshop, Portland, Or. (1987). 20 Id.; see also Pierson, supra note 1, at 6. 21 Wyllie, supra note 19. 22 Id.; see also Vanessa Bateman, Development of a Database to Man- age Rockfall Hazard: The Tennessee Rockfall Hazard Database, Pro- ceedings of the 82nd Annual Meeting of the Transportation Research Board, Washington, D.C. (2003). 23 F. Housley Carr, Preventing Rockslides Pays Off: State Officials Take Steps to Keep Debris from Falling on Nationâs Highways, 24 Engâg News-Rec., no.Â 28, Jul. 12, 1990 (quoting Wyllie that â[r]ockfall mitiga- tion efforts âhave really taken off â in the past three to five yearsâ). 24 Id.
6 NCHRP LRD 82 considered, the USMS originally required each slope to be eval- uated in 19 categories, for which each slope would be assigned a numeric score on a linear scale ranging from 1 (low hazard) to 10 (high hazard).45 The USMS introduced the concept of using weighting factors so that some categories could factor more heavily than others in the total hazard rating score.46 However, early sensitivity studies indicated that the weighting factors, and some of the categories, did not have a significant influence on the rankings of slopes based on hazard rating score.47 Fur- thermore, it was observed that the linear scoring scale made it difficult to differentiate between slopes, leading to the adoption of an exponential scoring scale.48 Over time, the USMS scoring system was revised to 11 categories, with each slope receiving a score of 3, 9, 27, or 81 for each category, and the 11 category scores then summed together to create the USMS slope hazard rating.49 The USMS is therefore similar to the RHRS in many respects. The USMS was originally envisioned as evolving into an expert system, which would recommend cost-efficient mitiga- tion strategies and repair schedules for slopes in the inventory.50 However, decisions to make capital improvements to a highway slope, or to defer remediation and risk failure of the slope, in- volve policy considerations that State DOTs have not ceded to software. Therefore, like the RHRS, the USMS remains a tool to assist decision-makers with identifying potential slope remedia- tion projects and analyzing the costs and benefits of undertak- ing such projects, but the USMS does not purport to make those decisions. 3. New York State Department of Transportation Rock Slope Rating Procedure Similar to the Oregon and Washington State DOTs, in 1988 the New York State DOT began to develop a statewide rating of rock slope locations based on Wyllieâs method.51 The New York State DOT modified the method in 1993 in an effort to distin- guish between the different types of categories and thus better risk assessment than failure mode.â) (citing S.Â Lowell & P.Â Morin, Unsta- ble Slope Management: Washington State, 207 TRÂ News at 11â15 (Mar.- Apr. 2000)). 45 Ho & Norton, supra note 41, at App.Â A. 46 Id. 47 Carlton L. Ho & Russell A. Knutson, Pilot Study of an Unstable Slope Management System, Washington State DOT Rep. No. WA-RDÂ 297.1, at 10â11, 27 (1994). 48 Id. at 11. 49 See, e.g., Woodard, supra note 6, at 14 (citing S.Â Lowell & P.Â Morin, Unstable Slope Management in Washington State, 207 TRÂ News at 11â15 (Mar.-Apr. 2000)). 50 Ho & Norton, supra note 41, at 26 (âUltimately, the system should be able to determine the optimal repair schedule. It should decide which slopes should be permanently fixed or if it is more eco- nomical to allow the site to fail each year.â). 51 New York State DOT, Rock Slope Rating Procedure, Geotechni- cal Engineering Manual GEM-15, 3 (2015), available at https:// www.dot.ny.gov/divisions/engineering/technical-services/technical- services-repository/GEM-15b.pdf [hereinafter GEM-15]. the highest rated slopes are among the most costly to repair.â35 Focusing limited construction and maintenance funds on the highest rated slopes, often described as a âworst-firstâ ap- proach, may not be the most cost-effective or rational use of those funds.36 The developers of the RHRS have long advocated that decision-makers consider a range of potential mitigation projects, and select projects based on both the cost of construc- tion and the reduction in rockfall hazard (based on how the project would impact the RHRS rating for a given site), per- haps using a benefit-cost analysis.37 On the other hand, focusing solely on projects with a favorable benefit-cost analysis could cause âsome of the highest rated slopes [to] be left unattended because of their higher cost.â38 The RHRS was not intended to be used to make those sorts of decisions regarding how to allocate limited slope stabilization resources, but rather to serve as a tool to assist the decision-makers. 2. Washington State Department of Transportation Unstable Slope Management System The Washington State DOT was one of the State DOTs that pooled funds for the development of the RHRS,39 and the Washington State DOT was an early adopter of the RHRS, using the Oregon DOT hazard rating system as early as 1988 to prepare priority rankings of rock slopes on a regional basis.40 Shortly thereafter, the Washington State DOT obtained FHWA grant funding to develop its own Unstable Slope Management System (USMS).41 The USMS was envisioned as a âmajor depar- tureâ from the RHRS.42 Notably, the USMS included soil slopes in addition to rock slopes and thus accounted for additional slope failure modes, such as landslides and shear failures of con- structed embankments.43 Like the RHRS, the USMS provided for slopes to be scored with respect to a number of categories, and the category scores were totaled to create an overall slope hazard rating. However, the scoring system for the USMS was initially very different from the RHRS. For one thing, the USMS developers elected to focus more heavily on categories related to the cost of a slope failure (e.g., the cost of repair and cleanup as well as the cost of property damage and personal injury) than factors related to slope failure potential.44 Also, in part because of the additional failure modes 35 RHRS Manual, supra note 29, at 76. 36 Kristen L. Sanford Bernhardt, Erik Loehr & Daniel Huaco, Asset Management Framework for Geotechnical Infrastructure, 9 J. Infra- structure Sys. 107, 109 (2003). 37 RHRS Manual, supra note 29, at 76â77. 38 Id. at 77. 39 Id. at 3. 40 Thomas C. Badger & Steve M. Lowell, Rockfall Control in Wash- ington State, 1343 Transp. Res. Rec.: J. Transp. Res. Bd. 14, 15 (1992). 41 Id. at 16; see also Carlton L. Ho & Sonja S. Norton, Develop- ment of an Unstable Slope Management System, Washington State DOT Rep. No. WA-RDÂ 270.1 (1991). 42 Bateman, supra note 22. 43 Badger & Lowell, supra note 40, at 16. 44 See, e.g., Woodard, supra note 6, at 14 (âThe developers of the USMS consider their system to be distinctive .Â .Â . with a greater focus on
NCHRP LRD 82 7 the recommendation of FHWA, and often with funding from FHWA.59 All unstable slope management programs derive to some extent from the rating system proposed by Wyllie and tend to be variations on the Oregon DOT RHRS, the Washington State DOT USMS, or the New York State DOT rock slope rating procedure. In conjunction with this study, the survey questionnaire contained in AppendixÂ B was sent to the 50 State DOTs, typi- cally to the office responsible for statewide asset management. Responses were received from 27 State DOTs, for a 54% re- sponse rate. Ten of the 27 respondents, or 37%, reported that the State DOT has adopted an unstable slope management program to provide a hazard priority ranking of unstable slopes, with other respondents indicating that unstable slope management programs are in various stages of planning and development. Many State DOTs have adopted the Oregon DOT RHRS more or less unmodified.60 However, a number of State DOTs using the RHRS have their own state-specific variations. For example, the âclimateâ category, related to freezing potential, was eliminated from the Tennessee DOT RHRS because of the relatively consistent climate across the state.61 The aver- age annual daily traffic (AADT) category was eliminated from the Montana DOT RHRS due to relatively low traffic volumes across the state.62 In the Utah DOT RHRS, the block size and rockfall history/frequency categories are weighted twice as heavily as other categories.63 The Oregon DOT and a number of other states using the RHRS have supplemented it with a land- slide rating system based largely on the Washington State DOT USMS.64 The Ohio DOT adopted both a rockfall hazard rating matrix and a landslide hazard rating matrix, both based on a limited subset of data categories due to the relatively flat geology in the state and the limited availability of data in some categories used by other StateÂ DOT programs.65 59 Tommy C. Hopkins, Tony L. Beckham, Liecheng Sun & Barry Butcher, Highway Rock Slope Management Program, Kentucky Transp. Center Rep. No. KTC-03-06/SPR-177-98-IF, 1â2 (2003). 60 Bateman, supra note 22. 61 Id. 62 Lawrence A. Pierson, Darren L. Beckstrand & Brent A. Black, Rockfall Hazard Classification and Mitigation Sys- tem, Rep. No. FHWA/MT-05-011/8174, 22 (2005). 63 Pack et al., supra note 53, at 50. 64 Id. at 16 (stating that the modified Oregon DOT RHRS âincludes landslides as well as rockfall. It is an adaptation of the Washington State DOTâs Unstable Slope Management System (USMS) .Â .Â ..â); Dan Pratt & Paul Santi, A Landslide Hazard Rating System for Colorado Highways, Proceedings of ASCE Rocky Mountain Geo-Conference, Lakewood, Colo. (2014). 65 See Robert Y. Liang, Landslide Hazard Rating Matrix and Data- base for the Ohio DOT, FHWA Rep. No. FHWA/OH-2007/18 (Dec. 2007); Woodard, supra note 6, at 14 (2004). estimate the risk associated with a rockfall site.52 In the New York State DOT rock slope rating procedure, seven categories related to geology (orientation of fractures, orientation of bed- ding planes, block size, rock friction, water/ice conditions, rock- fall history, and backslope above rock cut) are evaluated using the exponential scoring system and then summed together to calculate a Geologic Factor, which represents the relative risk of a rockfall occurring at the slope.53 Categories related to ditch/ shoulder effectiveness have been separated from the Geologic Factor categories, and the qualitative scores from the RHRS and USMS have been replaced with simple equations to calculate a Section Factor, which represents the relative risk that rocks will reach the pavement if a rockfall does occur.54 Finally, cat- egories related to vehicle exposure (such as sight distance and traffic volume) have been separated from the other categories and replaced with simple equations to calculate a Human Ex- posure Factor, which represents the risk of rocks impacting a vehicle if they do reach the pavement.55 Unlike the Geologic Factor, which retains the exponential scoring system of 1, 3, 9, or 81 corresponding to somewhat subjective determinations of severity, the Section Factor and Human Exposure Factor are objective (if simplistic) calculations based largely on geometric measurements.56 Rather than summing the scores for all geo- logic and geometric categories as in the RHRS and USMS scor- ing systems, the New York State DOT method multiplies the three risk scores (Geologic Factor, Section Factor, and Human Exposure Factor) to calculate the relative risk of personal injury or property damage to travelers due to rockfall at each slope.57 Although the New York State DOT rock slope rating procedure accounts for the same or very similar categories as those used in the RHRS and USMS, the New York State DOTâs method of combining the category scores into an overall hazard rating by multiplication rather than simple addition is advertised as better reflecting the risk to travelers at each site.58 4. Other State DOT Unstable Slope Management Programs The 21st century has seen the growing adoption of similar unstable slope management programs by other State DOTs to generate slope hazard ratings. The adoption of unstable slope management programs by individual State DOTs has come at 52 Id. 53 Id. at 4â5; see also Robert T. Pack, Ken Boie, Stoney Mather & Jamie Farrell, UDOT Rockfall Hazard Rating System: Final Report and Userâs Manual, Utah DOT Rep. No. UT-06.07, 18 (Jan. 2006). 54 GEM-15, supra note 51 at 4â5. 55 Id. at 5; see also Pack et al., supra note 53, at 18. 56 GEM-15, supra note 51, at 5, 8â9; see also Pack et al., supra note 53, at 49â50 (recognizing limitations of New York State DOT factor calculations). 57 GEM-15, supra note 51, at 4, App. B1. 58 Pack et al., supra note 53, at 18 (âThe factors used are similar to the [Oregon ]DOT system [RHRS] except they are formulated so as to be more directly related to the âprobabilitiesâ associated with conse- quences and hence risk.â).
8 NCHRP LRD 82 ï§ Quality (e.g., friction coefficient) o Differential erosion features and rates â¢ Presence and condition of mitigation measures (e.g., rock bolts, screens) â¢ Climate conditions o Water/rainfall conditions (e.g., annual precipitation) o Freeze/thaw conditions o Drainage features o Wind or seismic load conditions 3. Highway Data Data pertaining to the highway adjacent to the slope itself are generally related to the probability of debris reaching the pave- ment in the event of a slope failure.70 Unstable slope manage- ment programs typically include data categories to address most of the following, which can typically be directly measured as in the case of geometric dimensions: â¢ Highway functional class or type â¢ Pavement width and number of lanes â¢ Offset distance from slope â¢ Catchment ditch effectiveness o Ditch dimensions (depth and width) o Annual debris volume removed from ditch â¢ Typical impact of slope failure o Number of lanes impacted o Availability of detour route 4. Travel Data Data pertaining to vehicle travel on the highway are general- ly related to the probability of personal injury or property dam- age to travelers in the event that debris reaches the highway.71 Unstable slope management programs typically include data categories to address most of the following, which can typically be calculated or measured objectively: â¢ Sight distance and decision sight distance â¢ AADT â¢ Vehicle speed and/or speed limit â¢ Average vehicle risk (AVR) 5. Incident History Data With slope failures being difficult to predict scientifically, often the best indicator of future slope failures is the history of slope failure issues, including the average frequency of slope failures, the average volume of debris reaching the highway, and the average number of instances of personal injury or property damage when debris reaches the highway. Unstable slope man- agement programs typically include most of the following data categories: 70 See, e.g., id. 71 See, e.g., id. at 8. B. Data Almost all unstable slope management programs involve a database in which most of the following basic data types are stored for each slope in the inventory.66 1. Identification Data Data serving to identify the slopes in the inventory typically do not play a role in the calculation of the overall hazard rating. However, location data allow the sites to be displayed spatially using mapping software such as a Geographic Information Sys- tem (GIS). This can be useful for State DOTs to consider region- al remediation projects, such as relocation or realignment of a highway in a corridor that contains multiple hazardous slopes.67 Also, the programs can serve as a repository of photographs of slopes in the inventory, allowing geotechnical or geologic per- sonnel to consider whether a slope is degrading over time, and whether some of the other data categories should be adjusted to reflect the heightened hazard.68 Identification data that may be contained in the unstable slope management program include: â¢ Location o County or maintenance region o Highway number o Beginning and ending milepost of the slope o GPS coordinates â¢ Photographs (including dates) 2. Slope Data Data pertaining to the condition of the slope itself are gen- erally related to the likelihood of a slope failure occurring.69 Unstable slope management programs typically include data categories to address most of the following. Except for the slope geometry, many of these data items are not objective measure- ments but rather are scored subjectively (i.e., joint orientation may be deemed âfavorable,â ârandom,â or âadverseâ): â¢ Site geologic character (i.e., rock or soil) â¢ Slope dimensions o Height o Length along highway o Grade or steepness â¢ Structural condition o Joints or fractures ï§ Frequency ï§ Orientation 66 Scott L. Huang, Margaret m. Darrow & Peter Calvin, Unstable Slope Management Program, Rep. No.Â FHWA- AK-RD-12-14, at 23 (2009) (summarizing data categories used in the various unstable slope management programs). 67 See, e.g., Ho & Knutson, supra noteÂ 47, at 29â30. 68 See, e.g., Pierson, supra note 1, at 7. 69 See, e.g., GEM-15, supra note 51, at 5.
NCHRP LRD 82 9 The preliminary rating is heavily driven by local mainte- nance evaluations of slope failure potential and history.76 The preliminary ratings do not rely on significant data-gathering aside from questionnaires to be completed at the local main- tenance level.77 Preliminary ratings can thus be very subjective. Ideally, the local maintenance rating is corroborated or supple- mented by centralized State DOT personnel specially trained for the unstable slope management program, who would make observational visits to all sites that have been identified by local maintenance personnel as having significant potential hazard.78 This helps eliminate local bias and ensure consistency in pre- liminary ratings across the state. After the preliminary ratings are prepared for all slopes, the State DOT often elects to perform detailed ratings only for slopes falling into the high-to-moderate hazard groupings. 2. Detailed Rating Detailed ratings require collection and quantification of the data described in SectionÂ II.B. At minimum, this involves a field visit from centralized State DOT personnel with specialized training in the unstable slope management program to the sites that received high-to-moderate hazard preliminary ratings by the local maintenance personnel.79 The detailed ratings are composite scores, representing the combination of subscores for the individual categories of data that are deemed to contribute to the overall slope hazard. In most unstable slope management programs, the detailed rating is simply the sum of 10 to 12 individual category subscores.80 Each individual category subscore typically follows an expo- nential scale.81 For example, a taller slope results in an exponen- tially higher score in the âslope heightâ category, which is one component of the overall hazard rating score. The exponential scoring approach generates higher overall hazard rating scores for slopes that are extraordinary in one or more data categories, 76 See, e.g., Thomas C. Badger, Marc Fish & Tracy Trople, Manage- ment of Unstable Slopes Along Washington State HighwaysâPast, Pres- ent, and Future, Proceedings of GeoCongress 2013, San Diego, Cal., 1651 (2013). 77 See, e.g., Paul Thompson, Darren L. Beckstrand, Barry A. Benko, Aine E. Mines, Lawrence Pierson & Robert E. Kimmer- ling, Statewide Geotechnical Asset Management Program Development: Final Report for Rock Slopes, Unstable Soil Slopes and Embankments, Retaining Walls, Material Sites, AlaskaÂ DOT&PF Rep. No.Â STP4000(126) (A), at C-30, D-30 (2017). 78 See, e.g., Pierson, supra note 1, at 6 (âProper training in RHRS application is necessary to ensure the consistency of ratings between different raters.â). 79 See, e.g., RHRS Manual, supra note 29, at 9 (âBecause of proper training, they have since successfully demonstrated that reasonable and repeatable slope ratings can be achieved .Â .Â ..â); Huang, supra note 66, at 60 (âStaff training will promote accurate and reliable data collection for both new unstable slopes and updating existing slopes after the inven- tory has been completed.â). 80 Scott A. Anderson & Matthew J. DeMarco, Use of Rockfall Rating Systems in the Design of New Slopes, Proceedings of Biennial Geo- technical Seminar, Denver, Colo., 46â47 (Nov.Â 2012). 81 Huang et al., supra noteÂ 66, at 4, 30; Badger & Lowell, supra note 40, at 15; RHRS Manual, supra note 29, at 25â27. â¢ Typical slope failure mode (e.g., rockfall or landslide) â¢ Frequency of slope failures â¢ Magnitude of slope failures o Representative failure size (âblock sizeâ) o Annual debris volume removed from highway â¢ Maintenance frequency and/or annual maintenance costs â¢ Average cost of a slope failure event â¢ Annual number of instances of personal injury or prop- erty damage As unstable slope management programs were originally es- sentially âworst-firstâ programs used to identify slopes in need of full slope remediation, these data categories regarding fail- ure history could be âresetâ after a slope remediation project, causing the remediated slope to drop in the priority rankings. As programs transition from âworst-firstâ programs to geotech- nical asset management programs, the historical slope failure costs (e.g., of debris cleanup or other damage) can be used to determine the most cost-effective mitigation strategy, discussed in Section II.D. C. Hazard Rating Slope hazard rating scores generated by unstable slope man- agement programs help State DOTs identify the slopes that generally pose the greatest hazard and that are in need of at- tention.72 The initial identification of slopes to be included in the State DOTâs unstable slope management program typically involves collaboration between maintenance personnel at the local or regional level, and geotechnical or geologic personnel with more of a statewide focus. Out of the potential thousands of identified highway slopes per state, a two-stage rating process is typically employed to narrow the focus to slopes of concern.73 1. Preliminary Rating Typically a preliminary rating system is used to establish the inventory of slopes to be included in the unstable slope manage- ment program and to eliminate from consideration slopes with very low probability of failure or damage.74 The preliminary rat- ing establishes general groupings for the slopes, such as high, moderate, low, or no hazard, and typically does not attempt to distinguish between slopes within each grouping.75 72 See, e.g., Bigger Wall to Protect Road from Falling Rocks at âthe Dip,â Kodiak (AK) Daily Mirror, Feb.Â 20, 2014 (âThe transportation department said the Rezanof Drive project is a priority because the stateâs Unstable Slope Management Program recently categorized the Pillar Mountain hillside as the third most problematic hillside in the state based on hazards and the consequences of these hazards.â). 73 Huang et al., supra noteÂ 66, at 22 (âMost of the programs involve a two-stage implementation process, with preliminary and detailed evaluations.â). 74 See, e.g., Pierson, supra note 1, at 7. 75 See, e.g., RHRSÂ Manual, supra note 29, at 18.
10 NCHRP LRD 82 some form of benefit-cost decision-making support to better prioritize remediation projects.86 For a given slope, there may be a range of options that can be undertaken to mitigate the risk associated with slope failure, i.e., to reduce the likelihood of personal injury or property dam- age due to failure.87 Duncan Wyllieâs research in the 1990s and beyond has focused on determining the relative influence of the various data categories in contributing to rockfalls, and also the relative effectiveness of various rockfall mitigation measures.88 The following mitigation measures are listed in order of gener- ally increasing cost and effectiveness: â¢ Periodic inspections â¢ Continual monitoring (e.g., instrumentation) â¢ Warning signs or speed limit reductions â¢ Routine maintenance (e.g., cleaning out catchment ditch) â¢ Installation of barriers or guardrails â¢ Scaling and/or removal of material at risk of falling â¢ Installation of wire mesh, screens, or netting â¢ Drainage improvements (e.g., installation of horizontal pipe drains) â¢ Slope surface stabilization (e.g., with shotcrete or poly- urethane resin) â¢ Deep slope stabilization (e.g., with rock bolts or soil anchors) â¢ Construction of retaining walls â¢ Catchment ditch widening â¢ Highway widening â¢ Highway closure â¢ Highway relocation When unstable slope management programs were originally developed, some consideration was given to having the pro- grams ultimately prescribe mitigation measures such as repairs or maintenance schedules.89 In practice this has not occurred, as decisions regarding remediation involve balancing the risks and advantages of different remediation activities given lim- ited funds. State DOTs would prefer to have such remediation decisions made based on the judgment of experienced, skilled 86 See, e.g., Pierson, supra noteÂ 1, at 12; Badger et al., supra note 76, at 1650; Beckstrand et al., supra note 6, at 23. 87 See, e.g., Kyle Mittan, After Latest Slide, Is the Bluff Safe?, Aberdeen (WA) Daily World (Mar.Â 5, 2015) (describing various mitigation measures used on slopes in Washington State DOT USMS inventory). 88 See, e.g., Duncan C. Wyllie & Norman I. Norrish, Stabilization of Rock Slopes, in Landslides Investigation and Mitigation, ch. 18, TRBÂ Special Rep.Â 247, Transp. Res. Bd., Washington, D.C. (1996), at 474. 89 See, e.g., Ho & Norton, supra note 41, at 26 (âUltimately, the system should be able to determine the optimal repair schedule. It should decide which slopes should be permanently fixed or if it is more economical to allow the site to fail each year.â). helping to differentiate those slopes with one or more extraordi- nary characteristics even though the slopes may be ordinary in many or most data categories. In some unstable slope management programs, State DOTs have established unique or custom weighting factors to be ap- plied to some individual category subscores before they are summed together so that those categories have a greater or lesser influence on the detailed rating.82 Some unstable slope management programs follow the New York State DOT rat- ing procedure, in which subscores related to different risks are multiplied (rather than summed) together in an effort to better represent the relative risk of slope failure.83 The detailed rating is not a rigorous calculation or predic- tion of slope failure probability, which would require detailed and expensive computer modeling and simulation. Instead, it is a shorthand way to compare numerous slopes and identify those most worthy of more detailed consideration for remedia- tion or mitigation. After the detailed rating is prepared for all slopes, the State DOT often selects a threshold detailed rating score to further re- duce the set of slopes for which mitigation measures are consid- ered. For example, State DOTs may initially design slope reme- diation projects only for slopes with a detailed rating score above 500 (or some other threshold established by the State DOT). As the worst slopes are remediated, allowing the State DOT to direct resources toward less hazardous slopes, the threshold may be ad- justed accordingly.84 D. Mitigation The primary purpose of an unstable slope management pro- gram is to serve as a tool to help State DOTs make decisions regarding mitigation of slope hazards. Mitigation activities may range from routine maintenance (e.g., debris cleanup) to full remediation of the hazard (e.g., excavation of the slope or re- construction of the highway). The unstable slope management program typically does not prescribe mitigation activities that must be performed for unstable slopes. However, it can provide a framework for State DOTs to optimally allocate limited miti- gation funds to reduce slope failure hazards. The slope hazard ratings may serve as a starting point for remediation decisions, but it is widely understood that the âworst-firstâ approach to remediation (i.e., focusing remediation resources on the slopes with the highest hazard ratings) may not be the most rational way to allocate limited remediation funds, as the slopes with the highest hazard rating tend to also be the most expensive to re- mediate.85 Most unstable slope management programs include 82 Huang et al., supra noteÂ 66, at 3; Badger & Lowell, supra note 40, at 14; Woodard, supra note 6, at 194. 83 Huang et al., supra noteÂ 66, at 1. 84 Badger et al., supra note 76, at 1653â55. 85 Bernhardt et al., supra note 36, at 107, 109; David A. Stanley & Lawrence A. Pierson, Geotechnical Asset Management of Slopes: Condi- tion Indices and Performance Measures, Proceedings of GeoCongress 2013, San Diego, Cal., 1658, 1663 (Mar. 2013).
NCHRP LRD 82 11 which can be calculated from individual data that are included in the USMS and used to calculate the USMS hazard score.93 As in the RHRS method, the âcostâ or denominator of the USMS benefit-cost ratio is the cost of the remediation project. Recognizing the inherent limitation that the least expensive (and likely least effective) remediation strategies would other- wise tend to rise to the top of this project ranking system, the Washington State DOT advertises that only âall or nothingâ slope remediation projectsââcomprehensiveâ measures that will eliminate or significantly reduce the risk of slope failure for 20 yearsâare considered.94 As long as the âbenefitâ or numera- tor accurately reflects the 20-year cost of failing to remediate (e.g., 20 years of maintenance costs and traffic delay costs), then projects with a benefit-cost ratio of 1.0 or greater are those that are theoretically financially advantageous to the Washington State DOT. An obvious disadvantage of this approach is that it does not provide for consideration of mitigation measures short of âfull remediation,â which are probably less expensive and can provide for significant risk reduction, albeit for a period of time less than 20 years. Other drawbacks of the USMS approach include that the calculation of the âbenefitââi.e., the cost of doing nothingâ is approximate at best, as the simple calculation does not truly reflect the likely or expected cost of traffic delays due to slope failures over a 20-year period (as the probability of slope failure is not included in the calculation), and it does not account for any maintenance costs or traffic delay costs associated with the remediated slope (i.e., the full remediation project is presumed to be completely effective). With the advent of the Statewide Transportation Improve- ment Program (STIP) in Oregon, the Oregon DOT revised its priority ranking method to mirror the Washington State DOT USMS approach.95 In the Oregon DOTâs revised method, a modified RHRS score (based on the sum of only 5 of the origi- nal 12 RHRS factors) is multiplied by a maintenance benefit- cost factor, which is based on the ratio of the 20-year mainte- nance costs and the cost of the proposed repair.96 Finally, this number is multiplied by another factor based on the highway classification, so that a hazardous slope along an interstate high- way is weighted higher than one along a statewide highway, which in turn is weighted higher than a hazardous slope along a regional highway, etc.97 Again, a limitation of this method is that 93 Steve Lowell, William Gates, Lynn Moses, Chad Lukkarila, Brendan Fisher, Tom Badger & Norman Norrish, Conceptual Designs and Cost Estimates: A Critical Step in Managing Unstable Slopes Along Washington State Highways, Proceedings of 56th Highway Geology Symposium, Wilmington, N.C., 19, 27 (2005), available at http://www.highwaygeologysymposium.org/wp-content/uploads/56_ hgs-OPT.pdf. 94 Washington State DOT USMP, supra note 4, at 5. 95 Huang et al., supra note 66, at 7â8. 96 Id.; see also Curran Mohney, Unstable Slope Management for Oregon Highways, Proceedings of the Annual Meeting of the Portland Geological Society of America, Portland, Or., 7 (Oct.Â 2009), available at https://slideplayer.com/slide/6343077. 97 Huang et al., supra note 66, at 8. personnel and their managers.90 However, the data contained in unstable slope management programs, and the hazard ratings generated by those programs, can be used by skilled personnel to determine appropriate remedial measures by focusing reme- diation efforts on the categories that result in a high hazard rat- ing for a given slope. For example, where the hazard rating of a given slope is very high due to poor sight distance or ineffec- tive catchment ditch geometry, that could lead the State DOT to develop a conceptual design of a construction project focused on improving those elements of the highway. However, the con- ceptual design might be cost-prohibitive to construct, leading the State DOT to consider less expensive but also less effective mitigation strategies. 1. Benefit-Cost Analysis Most unstable slope management programs provide for State DOT engineers to develop one or more conceptual reme- diation projects for each slope, along with an estimated cost to construct or implement the conceptual design. In the survey performed in conjunction with this study, of the 10 State DOTs who reported using an unstable slope management program, six (60%) reported that the program is used to determine remedial or mitigation measures. The programs usually provide a mecha- nism to rank the remediation projects based on some form of benefit-cost analysis. For example, the developers of the Oregon DOT RHRS originally proposed ranking remediation projects by the ratio of the total RHRS rating score for the unremediated slope and the estimated cost of construction of the conceptual remedia- tion project.91 The highest priority projects would be those with the highest RHRS-to-cost ratio. The advantage of this project ranking approach is that very expensive remediation projects would only rise to a high priority ranking if the slope also had a very high RHRS rating. One obvious drawback with this project ranking approach is that it did not account for the effectiveness of the remediation project. If there were two conceptual projects for a given slope, the least expensive project would always have a higher priority than the more expensive project, regardless of their relative effectiveness in reducing slope hazards. The Washington State DOT USMS provided a project rank- ing based on a similar benefit-cost ratio approach, but modified to give the State DOT a better idea whether remediation makes financial sense.92 In the Washington State DOT USMS method, the âbenefitâ or numerator of the benefit-cost ratio is not the total USMS hazard rating score of the unremediated slope, but rather a calculation of the cost of failing to remediate, i.e., an- nual maintenance costs at the unremediated slope plus the traf- fic delay costs associated with cleaning up after a slope failure, 90 See, e.g., RHRS Manual, supra note 29, at 72 (âExperience is the best predictor of the effectiveness of a rockfall remedial design. There are too many gaps in our understanding of the mechanical properties of rock masses to rely wholly on an analytical approach.â). 91 Id. at 75â77. 92 Badger et al., supra note 76, at 1650, 1652.
12 NCHRP LRD 82 2. Evolution into Geotechnical Asset Management Programs In the 2012 transportation appropriations bill known as the Moving Ahead for Progress in the 21st Century Act (MAP-21),100 Congress enacted a requirement for all State DOTs to âdevelop a risk-based asset management plan for the National High- way System to improve or preserve the condition of the assets and the performance of the system.â101 âAsset managementâ is defined as a strategic and systematic process of operating, maintaining, and im- proving physical assets, with a focus on both engineering and eco- nomic analysis based upon quality information, to identify a struc- tured sequence of maintenance, preservation, repair, rehabilitation, and replacement actions that will achieve and sustain a desired state of good repair over the lifecycle of the assets at minimum practicable cost.102 Thus, an asset management plan should help State DOTs as- sess the condition, safety, and reliability of highway assets and rationally allocate funding for maintenance and construction projects based on risk to the highway system. Federal law cur- rently requires pavements and bridges in the National Highway System to be included in the State DOTâs asset management plan.103 However, Congress and FHWA have encouraged State DOTs to include in the asset management plan all other infra- structure assets (such as slopes, embankments, and retaining walls) within the right-of-way of the National Highway System and other public roads.104 Traditional unstable slope management programs such as the RHRS or the USMS are similar to asset management pro- grams in the sense that both types of programs âare intended to prioritize rehabilitation activities.â105 However, most unstable slope management programs are essentially âworst-firstâ rank- ing systems, aimed at prioritizing mitigation activities to pre- vent catastrophic slope failure.106 Asset management programs, on the other hand, consider the service life of an asset (such as a highway slope) and the costs of maintenance and failure over the assetâs life cycle in order to determine the most cost-effective maintenance strategy for the asset.107 The current trend is for ex- isting unstable slope management programs to be converted or integrated into geotechnical asset management programs that include natural soil slopes, rock slopes and faces, constructed embankments, and retaining walls along highways. The unsta- ble slope management programs serve as a âstarting pointâ for geotechnical asset management programs, supplemented with a more mature decision-making framework to determine the most cost-effective maintenance strategy for all slopes, not just 100 Pub. L. No. 112-141, 126 Stat. 405 (2012). 101 Id. Â§Â 1106(a) (codified at 23 U.S.C. Â§Â 119(e) (2019)). 102 23 U.S.C. Â§Â 101(a)(2) (2019). 103 23 U.S.C. Â§Â 119(e)(4)(A) (2019); 23 C.F.R. Â§Â 515.9(b) (2019). 104 23 U.S.C. Â§Â 119(e)(3) (2019); 23 C.F.R. Â§Â 515.9(c) (2019). 105 Bernhardt et al., supra note 36, at 107, 109. 106 Id. 107 Id. it effectively presumes that all remediation projects are equally effective at reducing risk of slope failure and eliminating annual maintenance costs, and it does not provide a way to compare alternative remediation measures for a given site based on rela- tive cost and effectiveness. The New York State DOT rock slope rating procedure, which, as discussed in Section II.B, attempts to more accurately assess the relative risk of personal injury or property damage asso ciated with a given slope, may lend itself to a more rational method of prioritizing remediation projects than either the RHRS or USMS methods.98 As discussed previously, the New York State DOT method groups the typical slope assessment scoring cate- gories into three subscoresâa Section Factor, a Geologic Factor, and a Human Exposure Factorâwhich are multiplied together to create an overall Total Relative Risk (TRR) score. A given slope may have a high TRR score due to both inadequate catch- ment ditch width (i.e., high Section Factor) and large, unstable rocks in the exposed face (i.e., high Geologic Factor). A con- ceptual remediation project to widen the catchment ditch could significantly improve the Section Factor component of the over- all TRR score, but it might be very expensive due to the large volume of rock that would have to be excavated. An alternative project to perform rock scaling and/or provide reinforcement in the form of rock bolts or netting could significantly improve the Geologic Factor component of the overall TRR score at a fraction of the price. A project that combines both efforts would be more expensive than either individual project but would sig- nificantly reduce both the Section Factor and Geologic Factor components of the TRR. The benefit-cost ratio for each project would be calculated as the anticipated reduction in the overall TRR associated with each remediation project, divided by the estimated cost of the remediation project. This method is con- sistent with the 1980s research of Wyllie, who recommended making remediation decisions by calculating the expected cost of a range of remediation approaches (e.g., no action, rock scal- ing, and rock bolting) based on the slope failure risk associated with each approach.99 The New York State DOT benefit-cost analysis method allows State DOTs to balance the risks and advantages of alter- native remediation approaches at a given slope. This benefit- cost calculation based on the anticipated reduction in TRR will naturally apply a higher weight to the more heavily traveled highways, as AADT factors into the Human Exposure Factor component of the overall TRR score. Therefore, the benefit-cost calculation provides a rational way to allocate limited funds to slope remediation across the New York State DOT slope inven- tory. As discussed in the following section, this approach is consistent with principles of asset management, particularly if life-cycle costs of alternative remediation measures are included in the analysis. 98 Pack et al., supra note 53, at 18â19. 99 Duncan C. Wyllie, N.R. McCammon & W. Brumond, Planning Slope Stabilization Programs by Using Decision Analysis, 749 Transp. Res. Rec.: J. Transp. Res. Bd. 34 (1980).