Implementation of the AASHTO Mechanistic-Empirical Pavement Design Guide and Software (2014)

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.

12 INTRODUCTION A survey was developed to determine the implementation efforts of U.S., Puerto Rico, and Canadian state highway and provincial transportation agencies in relation to the MEPDG and accompanying AASHTOWare Pavement ME Design™ software. The questionnaire focused on the practices, policies, and procedures that have been successfully used by highway agencies for implementing the MEPDG and AASHTOWare Pavement ME Design™. In addition, the survey requested information related to: • Agency decision-making authority for pavement design. • Organizational structure and steps required to work within this structure for successful implementation. • Use of consultants and in-house personnel for pavement design. • Level of staff expertise in ME pavement design principles. • Availability and quality of required data inputs. • Current status of implementation. • Reasons an agency has postponed or has yet to imple- ment the MEPDG and AASHTOWare Pavement ME Design™ software. • Agency implementation challenges or impediments. • Approaches and parties involved in the evaluation and adoption of the MEPDG. • Agency lessons learned that can be used to help other agencies in the implementation process. • Benefits accrued to the agency from implementation (tangible and intangible). • Development of training programs and implementation guides. The intended recipients of the survey questionnaire were the pavement design engineers of the state highway agencies, Puerto Rico, and the District of Columbia, and Canadian provincial and territorial governments. The detailed survey questionnaire is provided in Appendix A, and the agency responses are provided in Appendix B. As of March 2013, 57 agencies (90%) responded to the survey, including 47 U.S. highway agencies, Puerto Rico, and nine Canadian provincial and territorial governments. AGENCY PAVEMENT TYPES The following provides definitions used in the survey and in this synthesis for new construction pavement types, all of which are based on the pavement type definitions included in the MEPDG (AASHTO 2008). (Note: not all of the following pavement types have nationally calibrated pavement perfor- mance prediction models.) • Composite—new thin or thick asphalt surface layer over a new concrete layer. Base layers may consist of unbound aggregate and/or stabilized layers. • CRCP—concrete pavement with longitudinal reinforce- ment to hold shrinkage cracks tightly closed. Base layers may consist of unbound aggregate and/or stabilized layers. • Full-depth asphalt—relatively thick asphalt surface layer placed over stabilized subgrade or placed directly on subgrade. • JPCP—concrete pavement with short joint spacing, and with or without dowel bars (10 to 20 ft). Base layers may consist of unbound aggregate and/or stabilized layers. • Semi-rigid—thin or thick asphalt surface layer placed over a cementitious stabilized material. Base layers may consist of unbound aggregate and/or stabilized layers. • Thick asphalt—asphalt surface layer greater than 6 in. thick over unbound aggregate base layers. • Thin asphalt—hot or warm mix asphalt (that will be designated as asphalt in this synthesis, but is intended to imply either layer type) surface layer less than 6 in. thick placed over unbound aggregate base layers. In addition, the pavement type definitions for preservation and rehabilitation treatments used in the survey and in this synthesis include (not all of the following treatment types are included in or have nationally calibrated performance prediction models) (AASHTO 2008): • Bonded CRCP overlay—placing a CRCP overlay directly over (i.e., no interlayer) an existing concrete pavement that is in good structural condition. • Bonded JPCP overlay—placing a JPCP overlay directly over (i.e., no interlayer) an existing concrete pavement that is in good structural condition. • Cold in-place recycle (CIR) with asphalt overlay— milling (typically 3 to 4 in.) and mixing the existing asphalt surface with recycling agent, additives, and virgin aggregate, relaying, and compacting in-place followed by an asphalt overlay. • CIR without asphalt overlay—milling (typically 3 to 4 in.) and mixing the existing asphalt surface with recycling agent, additives, and virgin aggregate, relay- ing, and compacting in-place followed by either a thin chapter three SURVEY OF AGENCY PAVEMENT DESIGN PRACTICES

13 asphalt overlay and/or a chip seal(s) or other surface treatment(s). • Crack or break and seat with an unbonded overlay— crack or break and seat of an existing concrete pavement and overlay with an unbonded CRCP or JPCP overlay. • Crack or break and seat with asphalt overlay—crack or break and seat of an existing concrete pavement and overlay with an asphalt layer. • Dowel bar retrofit—placing dowel bars at the transverse joints and cracks of an existing JPCP concrete pavement to restore load transfer. • Diamond grinding—removing a thin layer (0.12 to 0.25 in.) of the existing concrete surface using equip- ment fitted with closely spaced diamond saw blades. • Full-depth reclamation (FDR) with asphalt overlay— removal of the full depth of the existing asphalt layer and predetermined portion of the underlying base by pulverizing, blending, and re-compacting followed by an asphalt overlay. • FDR without structural overlay—removal of the full- depth of the existing asphalt layer and predetermined portion of the underlying base by pulverizing, blending, and re-compacting followed by a thin asphalt overlay, chip seal(s), or other surface treatment(s). • Hot in-place recycle (HIR) with asphalt overlay— correction of distress within the upper 2 in. of an existing asphalt pavement by softening the asphalt surface layer with heat, mechanically loosening it, and mixing it with a recycling agent, unbound aggregates, rejuvenators, and/or virgin asphalt followed by an asphalt overlay. • HIR without asphalt overlay—correction of distress within the upper 2 in. of an existing asphalt pavement by softening the asphalt surface layer with heat, mechani- cally loosening it, and mixing it with a recycling agent, unbound aggregates, rejuvenators, and/or virgin asphalt followed by a surface treatment, thin asphalt overlay, or no treatment application. • Mill and asphalt overlay of existing composite—milling the surface of an existing composite pavement and over- laying with an asphalt overlay. • Mill and asphalt overlay of existing asphalt—milling the surface of an existing asphalt pavement and overlaying with an asphalt overlay • Rubblization with an unbonded overlay—fracturing an existing concrete pavement and overlaying with an unbonded concrete overlay. • Rubblization with asphalt overlay—fracturing an exist- ing concrete pavement and overlaying with an asphalt overlay. • Asphalt overlay of existing concrete—placing an asphalt overlay on an existing concrete pavement. • Asphalt overlay of existing asphalt—placing an asphalt overlay on an existing asphalt pavement. • Unbonded CRCP overlay—placing an interlayer (typi- cally asphalt) over an existing concrete pavement fol- lowed by placement of a CRCP overlay. • Unbonded JPCP overlay—placing an interlayer (typically asphalt) over an existing concrete pavement followed by placement of a JPCP overlay. Figure 1 provides a summary of responses on new construc- tion pavement types used by the responding agencies, including thick asphalt pavement (46 agencies), JPCP (44 agencies), thin asphalt pavement (41 agencies), and semi-rigid pavement (29 agencies). Agencies also indicated designing full-depth asphalt pavements (21 agencies) and composite pavements (18 agencies), and nine agencies reported designing CRCP. In addition, 12 agencies reported using other pavement types, FIGURE 1 Use of new construction pavement types.

14 with a chip seal(s) over unbound or bound aggregate layers as the predominant other pavement type (four agencies). Figures 2 and 3 summarize responses to the types of pres- ervation and rehabilitation treatments used by the responding agencies for concrete- and asphalt-surfaced pavements, respec- tively. For asphalt-surfaced pavement preservation and rehabil- itation treatments, 54 agencies use asphalt overlays of existing asphalt, 51 use mill and asphalt overlay of existing asphalt, and 42 use asphalt overlay of existing concrete pavements. Mean- while, 34 agencies indicated that they use FDR with an asphalt overlay, and 35 use mill and asphalt overlay of an existing composite pavement and rubblization with an asphalt overlay. For concrete-surfaced pavement preservation and rehabil- itation, the predominant treatment types included diamond grinding (44 agencies), dowel bar retrofit (34 agencies), and unbonded JPCP and CRCP overlays (27 agencies each). FIGURE 2 Use of concrete-surfaced preservation and rehabilitation pavement types. FIGURE 3 Use of asphalt-surfaced preservation and rehabilitation pavement types.

15 • Mississippi DOT—Dynatest ELMOD program (http:// www.dynatest.com/software/elmod) for asphalt overlay design of flexible and semi-rigid pavements. • Saskatchewan Highways and Infrastructure—Shell ME pavement design modified for Saskatchewan. • Texas DOT—Texas DOT CRCP–ME design method for concrete pavements (Ha et al. 2012) and the FPS21 for flexible pavement design (Liu and Scullion 2011). • Washington State DOT—Washington State DOT Ever- pave program (Mahoney et al. 1989) for asphalt overlay design. This information is further summarized in Figure 4 accord- ing to those agencies that use only empirical-based design procedures, empirical-based and MEPDG, MEPDG only, empirical-based and other ME design procedure, and only other ME design procedures. MECHANISTIC-EMPIRICAL PAVEMENT DESIGN GUIDE IMPLEMENTATION STATUS Three agencies reported that they have implemented the MEPDG, eight expressed no plans to implement the MEPDG at this time, 43 indicated that they plan to implement the MEPDG within five years, and three did not provide infor- mation on the timing of their implementation plans. Figure 5 provides a summary of the implementation status of the surveyed agencies. In a follow-up survey of responding agencies (conducted July 2013), many agencies indicated an on-going MEPDG implementation effort. For example, the following is a list of agency implementation activities: • Alabama—Currently concluding traffic study and have future plans for development of a materials library, followed by local calibration. • Arizona—Full implementation on major roadways is expected in early 2014. • Georgia—Currently conducting local calibration. AGENCY PAVEMENT DESIGN METHODS The transportation agencies surveyed currently use a variety of methods for pavement design, and most agencies (40) use more than one pavement design method for a given pavement type. AASHTO empirical methods are by far the most utilized, with 48 of the responding agencies using the AASHTO Interim Guide for Design of Pavement Structures (AASHTO 1972) through the AASHTO Guide for Design of Pavement Struc- tures, with 1998 Supplement (AASHTO 1998). Based on the results of the agency survey, the AASHTO 1993 Guide for the Design of Pavement Structures (AASHTO 1993) is the most commonly used design method, with 39 responding agencies reporting its use for at least one type of pavement design. Table 6 summarizes the agency pavement design methods. Twenty-four of the responding agencies mentioned the use of some type of ME design method. These methods include the MEPDG (used or being evaluated by 13 agencies) and other ME design methods developed by the agency or others (11 agencies). Three agencies have developed design cata- logs based on ME design procedures. The following agencies reported the use of these other ME design methods: • Alaska Department of Transportation and Public Facili- tates (ADOT&PF)—ME design procedure for asphalt pavements (ADOT&PF 2004). • California Department of Transportation (Caltrans)— CalME for flexible pavements (Ullidtz et al. 2010). • Colorado DOT—ME design procedure for bonded concrete overlays of asphalt pavements (Tarr et al. 1998). • Idaho Transportation Department—Winflex program (Bayomy 2006) or the Everpave program (Mahoney et al. 1989) for asphalt overlay design. • Illinois DOT—ME design procedure for flexible pave- ments, rigid pavements, and asphalt overlay of rubblized pavements (IDOT 2013). • Kentucky Transportation Cabinet—ME design process (Havens et al. 1981). • Minnesota DOT—MnPAVE for new flexible and asphalt overlay design. Method New Construction Rehabilitation Number of Agencies Asphalt Concrete Asphalt Concrete AASHTO 1972 7 2 5 1 7 AASHTO 1986 1 0 2 0 2 AASHTO 1993 35 23 31 19 39 AASHTO 1998 Supplement 4 11 4 8 13 AASHTO MEPDG1 12 10 10 7 13 Agency Empirical Procedure 7 1 9 3 13 WINPAS (ACPA 2012) 0 5 0 4 7 MS-1 (AI 1999) 1 0 3 0 3 ME-based Design Table or Catalog 1 3 0 2 3 Other ME Procedure 8 3 6 2 11 Other 5 7 7 8 14 1A number of agencies indicated that the MEPDG is currently being used or under evaluation; however, only three agencies indicated that the MEPDG has been implemented. TABLE 6 AGENCY USE OF PAVEMENT DESIGN METHODS

16 FIGURE 4 Agency pavement design methods. FIGURE 5 Summary of agency MEPDG implementation status.

17 pavement designs. The other two agencies, Missouri DOT and Oregon DOT, reported that they use empirical design methods in addition to the MEPDG. Of the 46 agencies that reported they are using or evaluat- ing the MEPDG, 45 indicated that it was being used or will be used to design new asphalt pavements, 39 that it was being used or will be used to design new JPCPs, and 12 that it was being used or will be used to design new CRCP. For over- lay thickness design, 38 agencies indicated that the MEPDG is being used or will be used to design asphalt overlays of existing asphalt pavements and 34 agencies are using or will use the MEPDG to design asphalt overlays of existing JPCP. Agencies also indicated that the MEPDG was being used or will be used to design asphalt overlays of fractured JPCP (27 agencies), unbonded JPCP overlays of existing JPCP (22 agencies), and JPCP overlays of existing asphalt pavements (21 agencies). A summary of agency MEPDG use by pavement type is shown in Table 7. Agencies that have not yet completed implementation stated that they needed to determine the benefits of using the MEPDG over their existing design method(s), develop an implemen- tation and training plan, and evaluate the applicability of the MEPDG to their current conditions. Less frequently cited was the need to obtain approval or buy-in from others in the agency or to evaluate the economic impacts of using the MEPDG method. MECHANISTIC-EMPIRICAL PAVEMENT DESIGN GUIDE CHAMPIONS Thirty-two of the responding agencies indicated that they have an MEPDG champion and 23 of the responding agencies indi- cated that the MEPDG champion is the state pavement engineer • Idaho—Consultant-conducted training on software oper- ation, and development of an ME user guide and imple- mentation roadmap. Internal staff is currently comparing design results between the current procedure and the MEPDG. Idaho Transportation Department is planning on using the current pavement design procedure as a start- ing point in the MEPDG once the performance prediction models have been locally calibrated; pavement designs will be required to meet the performance prediction crite- ria determined using the MEPDG. • Iowa—Locally calibrated the performance prediction models, but are currently re-evaluating the concrete performance prediction models. • Louisiana—Plans to begin the local calibration process, and has conducted comparisons between current proce- dure and the MEPDG on several interstate projects. • Michigan—Plans on transitioning to the MEPDG in 2014. • Mississippi—Performance prediction models are cur- rently being locally calibrated. Once the local calibra- tion has been completed, Mississippi DOT will conduct 2-year side-by-side comparison of results using the current procedure and the MEPDG. At this time, plan on implementing the MEPDG for the design of new or reconstructed pavements. • Oklahoma—MEPDG is being used for the design of new or reconstructed concrete pavements and concrete over- lays on interstate and other high-traffic routes. Oklahoma DOT is in the process of locally calibrating the asphalt pavement performance prediction models. • Ontario—Conducting local calibration with plans for implementation in 2014. • South Carolina—Conducting side-by-side comparisons and materials characterization. Future plans for local calibration. • Wisconsin—Completed studies related to asphalt mix- tures, concrete properties, and resilient modulus determi- nation of subgrade soils (http://wisdotresearch.wi.gov/ whrp). Wisconsin DOT is in the process of developing a user manual and conducting local calibration. Implemen- tation is anticipated to occur in 2014. Of the eight agencies that expressed no plans to implement the MEPDG, five are currently using agency-developed (or developed by others) ME and empirical design procedures. Five of the eight agencies reported that they consider their cur- rent design practices to be acceptable. Additional reasons cited by the eight agencies for not adopting the MEPDG at this time include software cost (four agencies), waiting for more agen- cies to implement the MEPDG (three agencies), and disagree- ment with the MEPDG modeling approach (two agencies). CURRENT AND EXPECTED USE OF THE MECHANISTIC-EMPIRICAL PAVEMENT DESIGN GUIDE Of the three agencies that noted having completed imple- mentation of the MEPDG, only one (Indiana DOT) reported that it uses the MEPDG exclusively for the evaluation of all TABLE 7 SUMMARY OF MEPDG USE OR PLANNED USE BY PAVEMENT TYPE Pavement Type Number of Responses New asphalt pavement 45 New JPCP 39 Asphalt overlay of existing asphalt pavement 38 Asphalt overlay of existing JPCP 34 Asphalt overlay of existing fractured JPCP 27 Unbonded JPCP overlay of existing JPCP 22 JPCP overlay of existing asphalt pavement 21 Asphalt overlay of existing CRCP 15 Bonded overlay of existing JPCP 13 New CRCP 12 Asphalt overlay of existing fractured CRCP 11 Unbonded JPCP overlay of existing CRCP 11 CRCP overlay of existing flexible pavement 7 Unbonded CRCP overlay of existing JPCP 7 Bonded concrete overlay of existing CRCP 6 Unbonded CRCP overlay of existing CRCP 6

18 IMPLEMENTATION CHALLENGES Agencies indicated that there were several challenges to imple- menting the MEPDG, including software complexity, avail- ability of needed data, defining input levels, and the need for local calibration. Software Agencies reported that AASHTOWare Pavement ME Design™ software is more complex than previous versions of AASHTO pavement design procedures. Agencies also indi- cated that training in ME fundamentals, MEPDG methods, and operation and functionality of the AASHTOWare Pave- ment ME Design™ software may be required. Data Availability As described previously, the MEPDG and AASHTOWare Pavement ME Design™ software requires significantly more data inputs than previous empirical and other developed ME pavement design software. Of the four types of data needed, agencies noted that pavement condition data (32 agencies) is the most readily available, followed by existing pavement structure data (31 agencies), and traffic data (28 agencies). Only 17 agencies indicated that materials data were readily available. In addition, agencies noted that obtaining materials characterization data required a significant level of effort to collect and additional equipment and field testing, as well as the additional time needed to conduct and evaluate the results and establish a materials library. Agencies also noted the challenge in obtaining traffic and materials data owing to agency divisional boundaries, and the unfamiliarity of other agency offices with the MEPDG data requirements and pave- ment design practices in general. INPUT LEVELS One of the features of the AASHTOWare Pavement ME Design™ software is its ability to use default, regional, or site- specific values for traffic and materials data inputs. Agencies reported that regional and site-specific data are needed for or pavement design engineer (or similar title or position). For 29 agencies, the MEPDG was or will be evaluated before implementation by the pavement design engineer, materials engineer, and pavement management, research, and design offices. The chief engineer (25 agencies) and pavement engi- neer (or similar title or position) (38 agencies) were listed as the most likely to ultimately decide whether or not the MEPDG should be implemented. AGENCY STRUCTURE As part of the agency survey, agencies were asked a number of organization-related questions. These included organiza- tional structure (centralized or decentralized), how effective communication was across agency functions (e.g., construc- tion, design, maintenance), which agencies had a MEPDG champion, and which agencies established a MEPDG over- sight committee. Table 8 provides a summary of the agency responses to these organizational questions. Although it is difficult to generate a direct relationship between an agency’s organizational structure and the MEPDG implementation status, doing so results in several interesting findings. For example, the agency structure does not appear to have an impact on the implementation status, but does indicate that most agen- cies (31 of the 44) responding to this survey question function under a centralized organizational structure (i.e., pavement designs are conducted, reviewed, and approved by the central or headquarters office). The communication level (consistent communication versus limited communication across agency functions) also does not appear to have much impact; there is a relatively even split across all implementation status, exclud- ing the three agencies that have implemented the MEPDG. However, all agencies that indicated that the MEPDG is or will be implemented within 2 years have an MEPDG champion (18 agencies). For agencies that indicated implementation will be more than 2 years, 14 indicated an MEPDG champion, whereas 12 agencies noted that they did not. In addition, the majority of agencies (five of six) that indicated that the MEPDG is or will be implemented within 1 year have an MEPDG oversight committee. For those agencies that indi- cated implementation will be greater than one year, 13 reported that they did have an MEPDG oversight committee, and 22 that they did not. Implementation Status Agency Structure (centralized/ decentralized) Communication Level1 (consistent/ limited) MEPDG Champion (yes/no) MEPDG Oversight Committee (yes/no) Implemented 2/1 3/0 3/0 2/1 Within 1 year 3/3 4/2 6/0 5/1 1 to 2 years 7/2 5/4 9/0 5/4 2 to 3 years 10/3 5/8 9/4 6/7 4 to 5 years 6/2 5/3 3/5 2/6 More than 5 years 3/2 3/2 2/3 0/5 1Consistent communication across agency functions; limited communication across agency functions. TABLE 8 SUMMARY OF AGENCY ORGANIZATION (NUMBER OF AGENCIES)

19 the MEPDG, but that it is expensive and time-consuming to collect. Vehicle classification and average annual daily truck traffic are the only site-specific traffic inputs that agencies are likely to have available for use. Many agencies do have regional inputs for hourly and monthly traffic adjustment factors and use default values for axles per truck, axle config- uration, wheelbase, and wander. Agencies stated that it requires additional time to compile all of the traffic data needed for regional and/or site-specific inputs. Table 9 provides a summary of responses related to the use of default, regional, and site-specific values for each of the traffic and material inputs. Of those agencies that responded to this survey question, the majority indicated the use of either the MEPDG default values or regional values. Relatively few agencies indicated the use of site-specific values. LOCAL CALIBRATION One of the steps in the local calibration process includes the evaluation and determination of how well the MEPDG predicted pavement performance (i.e., distress and IRI) cor- responds to observed field performance (AASHTO 2010). Traffic and Material Characteristic Number of Agencies MEPDG Regional Site-specific No response All Traffic 16 4 10 5 Vehicle class distribution 3 5 13 14 Hourly adjustment factors 12 12 9 2 Monthly adjustment factors 8 12 12 3 Axles per truck 8 14 9 4 Axle configuration 14 17 3 1 Lateral wander 15 16 4 0 Wheelbase 15 18 2 0 All Materials 24 1 7 3 All Asphalt Layers 22 2 8 3 Mixture volumetrics 10 3 16 6 Mechanical properties 11 7 12 5 Thermal properties 17 14 4 0 Asphalt Surface Layers Only 28 1 4 2 Mixture volumetrics 12 3 14 6 Mechanical properties 13 7 10 5 Thermal properties 20 12 3 0 Asphalt Base Layers Only 28 0 4 3 Mixture volumetrics 10 4 15 6 Mechanical properties 11 9 10 5 Thermal properties 21 11 3 0 All Concrete Layers 21 5 8 1 Poisson’s ratio 14 15 4 2 Unit weight 9 9 12 5 Thermal 16 13 3 3 Mix 11 4 12 8 Strength 11 4 13 7 All Chemically Stabilized Layers 25 4 6 0 Poisson’s ratio 21 9 2 3 Unit weight 21 5 6 3 Strength 19 4 8 4 Thermal 23 11 0 1 All Sandwiched Granular Layers 24 3 7 1 Poisson’s ratio 20 10 2 3 Unit weight 18 6 7 4 Strength 16 6 8 5 Thermal properties 22 10 1 2 All Non-stabilized Base Layers 23 3 6 3 Poisson’s ratio 13 17 3 2 Modulus 8 5 15 7 Sieve analysis 9 5 16 5 All Subgrade Layers 23 2 5 5 Poisson’s ratio 14 18 3 0 Modulus 5 5 16 9 Sieve analysis 6 7 14 8 All Bedrock Layers 20 6 3 6 Poisson’s ratio 20 12 1 2 Unit weight 21 2 10 2 Strength 20 10 2 3 TABLE 9 SUMMARY OF AGENCY INPUT VALUE USE

20 By making these comparisons, the agency is able to determine if local calibration of the performance prediction models is necessary or if the MEPDG performance prediction models are adequate. Because the MEPDG performance prediction models are based on data contained within the LTPP database, agency pavement condition measurements need to be consistent with the Distress Identification Manual (AASHTO 2010). At the time the Distress Identification Manual (Miller and Bellinger 2003) was being developed (circa 1987), transportation agen- cies were encouraged to adopt the standard distress definitions and to modify the procedures to fit their specific data collec- tion needs for pavement management and design. If the agency distress definitions are different than those used in the MEPDG (AASHTO 2008), the impact of the distress definition differ- ences needs to be evaluated (AASHTO 2010). Table 10 pro- vides a summary of agency survey responses in how well their distress definitions match those of the Distress Identification Manual. Most responding agencies indicated that IRI (40 agencies), rut depth (38 agencies), and alligator cracking (36 agencies) data were consistent with the Distress Identification Manual. For JPCP, the responding agencies reported that transverse cracking (35 agencies) and faulting (33 agencies) were con- sistent with the LTPP method, as was punchout measurement for agencies that constructed CRCP. The most often reported distresses that were not consistent with LTPP data collection procedures include longitudinal cracking, thermal cracking, and reflective cracking for asphalt pavements. A number of agencies indicated that they have conducted local calibration of the asphalt and/or concrete performance prediction models contained within the MEPDG. Five agencies indicated local calibration of the asphalt models (Table 11) and seven indicated calibration of the concrete models (Table 12). In addition, the Hawaii and New Jersey DOTs reported that the asphalt IRI performance prediction model has been cali- brated to local conditions; however, the remainder of the asphalt pavement performance models will be conducted at a future date. During the local calibration process an agency defines the threshold limits and reliability limits for each of the perfor- mance prediction models. Those agencies that indicated local calibration of the performance prediction models had been conducted were asked to provide the threshold limits, reli- ability levels, and model coefficients for each locally cali- brated performance prediction model. Tables 13 and 14 list the MEPDG default performance criteria and reliability levels for asphalt and concrete pavement, respectively. Tables 15–21 provide the performance criteria and reliability levels for the Colorado, Florida, and Arizona DOTs. Information on perfor- Pavement Type Condition Indicator Did Not Respond Distress Not Used Agency Distress Definition Similar to LTPP Agency Distress Definition Not Similar to LTPP All IRI 12 2 40 3 Asphalt Longitudinal cracking 15 7 32 3 Alligator cracking 15 0 36 6 Thermal cracking 17 4 28 8 Reflective cracking 15 6 30 6 Rut depth 14 0 38 5 JPCP Transverse cracking 17 2 35 3 Joint faulting 17 4 33 4 CRCP Punchouts 38 6 11 2 TABLE 10 AGENCY DISTRESS DEFINITIONS (NUMBER OF AGENCIES) Agency IRI Longitudinal Cracking Alligator Cracking Thermal Cracking Rut Depth Reflective Cracking Asphalt layer Total Arizona Do not use MEPDG Colorado Hawaii 1 1 1 1 1 1 Indiana Do not use MEPDG MEPDG Do not use Do not use Missouri MEPDG MEPDG MEPDG New Jersey 1 1 1 1 1 1 Oregon MEPDG 1Future plans. Indicates performance prediction models have been locally calibrated. TABLE 11 AGENCY LOCAL CALIBRATION—ASPHALT MODELS

21 mance threshold limits and reliability levels was also obtained for the Indiana, Missouri, and Oregon DOTs. The performance threshold limits and reliability levels for these three agencies are included as part of the agency case examples described in chapter five. Tables 15 and 16 provide the Colorado DOT’s perfor- mance threshold limits and reliability levels for asphalt and concrete pavements, respectively. The Colorado DOT distress threshold limits were determined from pavement management data, whereas the reliability levels were based on input from pavement managers and pavement design staff. Florida DOT has locally calibrated the JPCP performance prediction models. Threshold limits and reliability levels used by Florida DOT are provided in Table 17. The performance threshold limits and reliability levels are based on existing practice and experience, ranges provided in the MEPDG (AASHTO 2008), and engineering judgment. Arizona DOT has locally calibrated both asphalt and JPCP performance prediction models. Threshold limits for asphalt and concrete pavements are provided in Tables 18 and 19, respectively. Table 20 provides the reliability levels, by functional class, used by Arizona DOT. Performance thresholds and reliability levels are based on engineering judgment, pavement management criteria, sensitivity analy- sis, functional class (i.e., a higher reliability to minimize the consequence of early failure on more heavily trafficked routes), and previous Arizona DOT reliability levels. The threshold limit for IRI is based on the Arizona DOT standard speci- fications and values achieved regularly during construction (see Table 21). Tables 22 and 23 provide a summary of the agency cali- bration coefficients for concrete and asphalt pavements, respectively. Note shaded cells indicate that the agency uses a different calibration coefficient value than the MEPDG default value. It can also be noted that there is significant variation between the agency reported and MEPDG default calibration coefficients, in some instances the difference is an order of magnitude. Agency JPCP CRCP IRI Transverse cracking Faulting IRI Punchouts Arizona Colorado Do not use Do not use Florida Do not use Do not use Indiana MEPDG MEPDG Do not use Do not use Missouri MEPDG MEPDG Do not use Do not use North Dakota MEPDG MEPDG Do not use Do not use Oregon MEPDG MEPDG Indicates performance prediction models have been locally calibrated. TABLE 12 AGENCY LOCAL CALIBRATION—CONCRETE MODELS Performance Criteria Limit Reliability Initial IRI (inches/mile) 63 N/A1 Terminal IRI (inches/mile) 172 90 Longitudinal cracking (feet/mile) 2,000 90 Alligator cracking (percent area) 25 90 Transverse cracking (feet/mile) 250 90 Chemically stabilized layer—fatigue fracture (percent) 25 90 Permanent deformation—total pavement (inches) 0.75 90 Permanent deformation—asphalt only (inches) 0.25 90 Reflective cracking (percent) 100 501 Asphalt overlay—JPCP transverse cracking (percent slabs) 15 90 Asphalt overlay—CRCP punchouts (number per mile) 10 90 N/A = not available. 1Cannot be changed by the user. TABLE 13 MEPDG DEFAULT CRITERIA AND RELIABILITY VALUES—ASPHALT TABLE 14 MEPDG DEFAULT CRITERIA AND RELIABILITY VALUES—CONCRETE Performance Criteria Limit Reliability Initial IRI (in./mi) 63 N/A1 Terminal IRI (in./mi) 172 90 JPCP transverse cracking (percent slabs) 15 90 JPCP mean joint faulting (in.) 0.12 90 CRCP punchouts (number per mile) 10 90 N/A = not available. 1Cannot be changed by the user.

22 Functional Class IRI 1, 2 (in./mi) Transverse Cracking1 (percent slabs) Mean Joint Faulting3 (in.) Reliability (percent) Interstate 160 7.0 0.12 80–95 Principal Arterial 200 7.04 0.14 75–95 Minor Arterial 200 7.04 0.20 70–95 Major Collector 200 7.04 0.20 70–90 Minor Collector 5 5 5 50–90 Local 5 5 5 50–80 1New construction, determines the year to first rehabilitation (minimum age of 27 years). 2Rehabilitation, maximum value at end of design life. 3Maximum value at end of design life. 4Under evaluation. 5To be determined. TABLE 16 PERFORMANCE CRITERIA AND RELIABILITY (CONCRETE)— COLORADO DOT Functional Class IRI (in./mi) Transverse Cracking (percent slabs) Mean Joint Faulting (in.) Reliability (percent) All 180 10 0.12 75–95 TABLE 17 PERFORMANCE CRITERIA AND RELIABILITY (CONCRETE)—FLORIDA DOT Functional Class IRI (in./mi) Alligator Cracking (percent area) Transverse Cracking (ft/mi) Total Rut Depth1 (in.) Total Cracking2, (percent area) Interstate 150 10 1,000 0.50 10 Primary 150 15 1,500 0.50 15 Secondary 150 25 1,500 25 1At the end of a 15-year performance period. 2Alligator + reflective. TABLE 18 PERFORMANCE CRITERIA (ASPHALT)—ARIZONA DOT Functional Class IRI (in./mi) Mean Joint Faulting (in.) Transverse Cracking (percent slabs) Interstate 150 0.12 10 Primary 150 0.12 15 Secondary 150 0.12 25 TABLE 19 PERFORMANCE CRITERIA (CONCRETE)—ARIZONA DOT Functional Class Reliability (percent) Interstate and Freeway 97 Non-interstate Highways (>10,000 ADT) 95 Non-interstate Highways (2,001 to 10,000 ADT) 90 Non-interstate Highways (501 to 2,000 ADT) 80 Non-interstate Highways (<500 ADT) 75 ADT = average daily traffic. TABLE 20 RELIABILITY LEVELS—ARIZONA DOT TABLE 15 PERFORMANCE CRITERIA AND RELIABILITY (ASPHALT)—COLORADO DOT Functional Class IRI 1, 2 (in./mi) Longitudinal Cracking1, 2 (ft/mi) Alligator Cracking2, 3 (percent area) Transverse Cracking2, 3 (ft/mi) Asphalt Rut Depth1, 2 (in.) Total Rut Depth1, 2 (in.) Total Cracking3, 4 (percent area) Reliability (percent) Interstate 160 2,000 10 1,500 0.25 0.40 55 80–95 Principal Arterial 200 2,500 25 1,500 0.35 0.50 105 75–95 Minor Arterial 200 3,000 35 1,500 0.50 0.65 155 70–95 Major Collector 200 3,000 35 1,500 0.50 0.65 155 70–90 Minor Collector 5 5 5 5 5 Local 5 5 5 5 5 5 5 5 5 50–80 50–90 1New construction, determines the year to first rehabilitation (minimum age of 12 years). 2Rehabilitation, maximum value at end of design life. 3Maximum value at end of design life. 4Alligator + reflective. 5To be determined.

23 TABLE 21 INITIAL IRI VALUES—ARIZONA DOT Pavement Type Initial IRI (in./mi) New and Reconstructed Asphalt 45 Asphalt Overlay of Existing Asphalt Pavement 52 New JPCP 63 Asphalt Rubber Friction Course over JPCP or CRCP 50 Feature MEPDG Arizona Colorado Florida Missouri Cracking C1 2.0 2.0 2.0 2.8389 2.0 C2 1.22 1.22 1.22 0.9647 1.22 C4 1.0 0.19 0.6 0.5640 1.0 C5 –1.98 –2.067 –2.05 –0.5946 –1.98 Std. Dev. 1 4 7 1 1 Faulting C1 1.0184 0.0355 0.5104 4.0472 1.0184 C2 0.91656 0.1147 0.00838 0.91656 0.91656 C3 0.002848 0.00436 0.00147 0.002848 0.002848 C4 0.000883739 1.1E-07 0.008345 0.000883739 0.000883739 C5 250 20000 5999 250 250 C6 0.4 2.309 0.8404 0.0790 0.4 C7 1.8331 0.189 5.9293 1.8331 1.8331 C8 400 400 8 400 400 Std. Dev. 2 5 2 2 Punchout Not applicable Not applicable C1 2.0 2.0 2.0 C2 1.22 1.22 1.22 C3 216.8421 85 216.8421 C4 33.15789 1.4149 33.15789 C5 –0.58947 –0.8061 –0.58947 Crack Std. Dev. 3 6 3 IRI (CRCP) Not applicable Not applicable C1 3.15 3.15 3.15 C2 28.35 28.35 28.35 Std. Dev. 5.4 5.4 5.4 IRI (JPCP) J1 0.8203 0.6 0.8203 0.8203 0.82 J2 0.4417 3.48 0.4417 0.4417 1.17 J3 1.4929 1.22 1.4929 2.2555 1.43 J4 25.24 45.2 25.24 25.24 66.8 Std. Dev. 5.4 5.4 5.4 5.4 5.4 7Pow(57.08 x CRACK, 0.33) + 1.5 80.0831 x Pow(FAULT,0.3426) + 0.00521 90.1 for A-7-6 soils 100.001 for A-7-6 soils 113 for A-7-6 soils 1Pow(5.3116 x CRACK,0.3903) + 2.99 2Pow(0.0097 x FAULT,0.05178) + 0.014 32 + 2.2593 x Pow(0.4882 x PO) 4Pow(9.87x CRACK,0.4012) + 0.5 5Pow(0.037 x FAULT,0.6532) + 0.001 61.5 + 2.9622 x Pow(PO,0.4356) 1 1 1 TABLE 22 AGENCY LOCAL CALIBRATION COEFFICIENTS—CONCRETE ACTIVITIES TO AID IMPLEMENTATION Agencies were asked about specific activities that might help them in implementing the MEPDG. The following is a sum- mary of the responses: • Training in AASHTOWare Pavement ME Design™ (35 agencies). • Assistance with calibrating models to local conditions (35 agencies). • Dedicated AASHTO MEPDG/ME Design website for sharing technical information (34 agencies). • Training in interpretation of AASHTO Pavement ME Design software results (32 agencies). • Training in methodology for obtaining AASHTO MEPDG/ME Design inputs (31 agencies). • Training in ME design principles (28 agencies). • Training in how to modify pavement sections to meet design criteria (25 agencies). • Establishment of an expert task or user group (24 agencies). • Ability to share ME Design databases with other agencies (17 agencies). CHALLENGES AND LESSONS LEARNED Survey respondents provided a number of challenges and lessons learned during the implementation process. One of the more common responses was the lack of readily available

24 Feature MEPDG Arizona Colorado Missouri Oregon Cracking C1 Bottom 1.0 1.0 0.07 1.0 0.56 C1 Top 7.0 7.0 7.0 7.0 1.453 C2 Bottom 1.0 4.5 2.35 1.0 0.225 C2 Top 3.5 3.5 3.5 3.5 0.097 C3 Bottom 6000 6000 6000 6000 6000 C3 Top 0 0 0 0 0 C4 Top 1000 1000 1000 1000 1000 Std. Dev. Top 1 1 1 Std. Dev. Bottom 2 2 12 2 2 1 1 Fatigue BF1 1 249.00872 130.3674 1 1 BF2 1 1 1 1 1 BF3 1 1.23341 1.2178 1 1 Thermal Fracture Level 1 1.5 1.5 7.5 0.625 1.5 Level 2 0.5 0.5 0.5 0.5 0.5 Level 3 1.5 1.5 1.5 1.5 1.5 Std. Dev. (Level 1) 3 3 3 3 3 Std. Dev. (Level 2) 4 4 4 4 4 Std. Dev. (Level 3) 5 5 5 5 5 Rutting (asphalt) BR1 1.0 0.69 1.3413 1.48 BR2 1.0 1.0 1.0 1.0 BR3 1.0 1.0 1.0 0.9 Std. Dev. 6 9 14 6 6 Rutting (subgrade) BS1 (fine) 1.0 0.37 0.84 0.4375 1.0 Std. Dev. (fine) 7 10 15 7 7 BS1 (granular) 1.0 0.14 0.4 0.01 1.0 Std. Dev. (granular) 8 11 16 8 8 IRI C1 (asphalt) 40 1.2281 35 17.7 40 C2 (asphalt) 0.4 0.1175 0.3 0.975 0.4 C3 (asphalt) 0.008 0.008 0.02 0.008 0.008 C4 (asphalt) 0.015 0.028 0.019 0.01 0.015 C1 (over concrete) 40.8 40.8 40.8 40.8 40.8 C2 (over concrete) 0.575 0.575 0.575 0.575 0.575 C3 (over concrete) 0.0014 0.0014 0.0014 0.0014 0.0014 C4 (over concrete) 0.00825 0.00825 0.00825 0.00825 0.00825 1200 + 2300/(1 + exp(1.072 – 2.1654 x LOG10 (TOP + 0.0001))) 110.05 x Pow(BASERUT, 0.115) + 0.00110 21.13 + 13/(1 = exp(7.57 – 15.5 x LOG10(BOTTOM + 0.0001))) 121 + 15(1 + exp(-1.6673 – 2.4656*LOG10(BOTTOM+0.0001))) 30.1468 x THERMAL + 65.027 13Under review 40.2841 x THERMAL + 55.462 140.2052 x Pow(RUT,0.4) + 0.001 50.3972 x THERMAL + 20.422 150.1822 x Pow(SUBRUT, 0.5) +0.001 60.24 x Pow(RUT, 0.8026) + 0.001 160.2472 x Pow(BASERUT, 0.67) + 0.001 70.1235 x Pow(SUBRUT, 0.5012) + 0.001 170.01 for A-7-6 soil 80.1447 x Pow(BASERUT, 0.6711) + 0.001 90.0999 x Pow(RUT, 0.174)+0.001 100.05 x Pow(SUBRUT, 0.085) + 0.001 TABLE 23 AGENCY LOCAL CALIBRATION COEFFICIENTS—ASPHALT

25 traffic and materials data, and the large effort required to obtain the needed data. Agencies also indicated that contracting the applicable office (e.g., materials, traffic) early on in the implementation process to make sure that everyone understands what data are needed and why, and being prepared to conduct field sampling and testing if the needed data are not available. The following summarizes the responses: • Challenges (one agency response for each statement) – District offices are resistant to change from empirical- based designs to ME-based designs. There is a higher comfort level with the inputs and resulting outputs (i.e., layer thickness) with the AASHTO 1993 Guide. Making the shift to using design inputs and predicting distresses in the MEPDG, rather than obtaining layer thickness as the final result, has been difficult. – Changes to the pavement condition data collection procedures that have resulted in inconsistency with data measurement and the ability to obtain reliable pavement condition data for use in the calibration process. – Lack of resources to conduct local calibration and training of staff. – The MEPDG is too complex for most practicing engineers; however, this may be improved through training to increase the engineer’s confidence in the design procedure. – Rework required as a result of newer versions of soft- ware that yield different results than previous versions (e.g., moving from NCHRP 1-37A to MEPDG v1.1) and the difference required recalibration of perfor- mance prediction models to local conditions. • Lessons learned (one agency response for each statement) – Establish realistic timelines for the calibration and validation process. – Allow sufficient time for obtaining materials and traffic data. – Ensure the data related to the existing pavement layer, materials properties, and traffic is readily available. – If necessary, develop a plan for collecting the needed data; this can require an expensive field sampling and testing effort. – Develop agency-based design inputs to avoid varying inputs and outputs to minimize design variability. – Provide training to agency staff in ME design funda- mentals, MEPDG procedures, and the AASHTOWare Pavement ME DesignTM software.