Trip Generation Rates for Transportation Impact Analyses of Infill Developments (2013)

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.

11 This chapter presents analysis used by the research team to select an approach to estimating infill trip generation. This approach serves as the foundation for the detailed estimation methods developed in this research study. The review of the methods presented in Table 2.2 estab- lished an understanding of the features, requirements, and limitations that may be expected from estimation methods developed under each of the broader approach categories. This understanding was used as input to develop the criteria for selecting an approach for estimating infill trip generation and as input into the development of the proposed methods described in Chapter 4. The selection of an approach has fundamental implica- tions for the development of an infill trip generation method, including considerations such as the amount of initial data required, the level of effort the practitioner may need to invest in using the method, and other significant attributes that affect the viability of the resulting proposed method. 3.1 Basis for Selecting an Approach Evaluation criteria were established to determine which of the approach categories would best meet the research objec- tive of this study. The evaluation criteria generally fall into one of the following areas: The ability to produce required information for use in traffic impact analyses. This group of criteria seeks to ensure that the proposed method produces data that conform to the current practice in site and traffic impact analysis. At a mini- mum, the practice requires any trip generation method to estimate morning and afternoon peak hour directional vehi- cle trips for a given land use category (LUC) using commonly available units of development such as building floor area, employees, or dwelling units (independent variables). So that the proposed methodology can meet the needs of emerging multimodal impact analysis techniques, it should use person trips as the common denominator when converting between conventional ITE vehicle-trip data and infill trip data and as a basis for determining peak hour mode share of infill develop- ment. This approach will also support emerging measures of effectiveness that define capacity in terms of person through- put and that consider multiple modes rather than automobile throughput. Ease of use. This group of criteria assesses the ease of using the proposed method by the practitioner. A method’s sim- plicity and convenience promotes its continued use as a tool and its eventual adoption by practitioners and agencies as the state of the practice. The criteria used to evaluate ease of use by practitioners favor approaches that are intuitive, expand on but do not replace established methodologies, and use data familiar to the practitioner that are locally attainable. Credibility of an approach in terms of its reliability and consistency of performance. This criterion gives credence to approaches that give consistent and repeatable results while remaining sensitive to variations in factors that the user expects to result in a different outcome. The level of effort required to further develop or expand methodologies within an approach category to fully meet the research objective. These criteria are intended to deter- mine the likely feasibility of methodologies based on the level of effort that it will take to expand/develop the methods to meet the other requirements. For example, a method based on an approach that directly estimates infill trip generation using multivariate regression equations, with readily available input data, and that is vali- dated to accurately predict trips at a very high level of con- fidence is a promising method. Yet, if this method has only been developed for one LUC and has only been validated in one county of the United States, it is not a practical nor use- ful method for the profession in general and would likely require substantial development to be universally applicable. The level of effort and cost to fully develop an approach into a method with widespread usability is a critical determinant of its feasibility. C H A P T E R 3 Research Approach

12 The criteria considered by the research team in identifying approaches that result in estimation methods with long-term potential to serve the profession are: 1. The approach is compatible with existing traffic impact analysis methods (i.e., ability to estimate peak hour, directional-dependent variables). 2. The approach uses person trips as the common denomi- nator when converting between baseline ITE vehicle-trip data and infill trip data, and as a basis for determining peak hour mode share of infill development. 3. The approach is adaptable to meet the needs of emerging multimodal impact analysis techniques. 4. Input data needed to apply the approach are readily avail- able, or the ease and cost of collecting and applying the data are reasonable. 5. The approach requires no specialized skills or invest- ment in training or software in order to estimate infill trip generation. 6. The approach applies to the land uses in the ITE Trip Generation Manual and has few, if any, restrictions on land use categories and geography. 7. Computations used in the approach are intuitive to the user and transparent by way of documentation. 8. There is a reasonable cost for collecting data necessary to develop and validate the methodology. 9. The approach has the ability to take advantage of existing trip generation databases or other readily available data used as input to the method. 10. The principles of development and application of the approach can be described to laypeople. 11. The approach would have likely acceptance by members of the transportation planning and traffic engineering pro- fession who prepare and review site traffic impact analyses. 12. The approach allows for easy addition to or expansion of the database (i.e., new data points or new land uses) once original database is established. Table 3.1 compares the candidate approaches against the selection criteria. 3.2 Selection of an Approach and Conclusion The research team selected an approach on which to base the proposed methodology for estimating infill trip genera- tion: (a) use the research objective and associated selection criteria presented previously, (b) review recent research on infill trip generation estimation methods, and (c) use the research team’s collective experience in estimating trip gen- eration for numerous types of development and preparing site-specific transportation impact analyses. The research team selected the approach of using ITE rate adjustment based on empirical data, given that the approach met the research objective and, to varying degrees, all of the selection criteria. Specifically, some of the more critical con- siderations that drove the selection were: • The approach has compatibility with existing traffic impact methods. The selected approach is essentially an enhancement of the trip generation step in the well- established four-step transportation modeling process (trip generation, trip distribution, mode choice, and trip assignment). The approach plugs in to the standard methods practitioners already use, so the learning curve is minimal. • Input data needed are readily available, or the effort and cost of collecting and applying the data are reasonable. The data required, and the methods of acquiring it, are familiar to the practitioner who prepares traffic impact analyses. The data collection procedures in ITE’s guidance on conduct- ing trip generation studies (3) are sufficient. The area where users may need assistance is in selecting data collection sites in the appropriate context—a topic addressed in this report. • The approach uses person trips as the common denomi- nator when converting between baseline ITE vehicle-trip data and infill trip data. Through simple conversion, the use of person trips to derive vehicle trips and estimate the pedestrian, bicycle, and transit passenger trip generation of infill development significantly increases the robustness of the site travel characteristics and better informs the pro- cess of selecting appropriate transportation solutions than through conventional impact analyses. • The approach applies to the land uses in the ITE Trip Gen- eration Manual and has few, if any, restrictions on land use categories and applicable geography. The approach of employing empirical data provides the practitioner with maximum flexibility in that there are no limitations or con- straints in regard to land use classification or geography. Even the step of collecting data at similar sites is consistent with the conventional impact analysis process (e.g., field collection of vehicular turning movement data, site reconnaissance). • It is easy to add to or expand the database (i.e., new data points or new land uses) once original database is estab- lished. This approach does not require expansion of the existing ITE trip generation database or the need to add land use classifications. Practitioners may choose to estab- lish a new database of context-sensitive person-trip and mode-share data for specific land uses, including a uniform method of classifying context. While the research panel supported the selection of the ITE rate adjustment approach using empirical data from proxy sites, it also encouraged exploration of an alternate method

13 Table 3.1. Comparison of approach categories against selection criteria. (continued on next page) Criteria Grouping Criteria Direct Estimation Based on Regression Analysis Direct Estimation Based on Empirical Data ITE Rate Adjustment Based on Regression Analysis ITE Rate Adjustment Based on Empirical Data Extracted from HTSs Collected at Proxy Sites A bi lit y to p ro du ce re qu ire d in fo rm at io n fo r u se in e st ab lis he d tr af fic im pa ct a n al ys es p ro ce du re s (1) Approach is compatible with existing traffic impact analysis methods (i.e., ability to estimate peak hour, directional- dependent variables). (2) Approach uses person trips as the common denominator when converting between baseline ITE vehicle-trip data and infill trip data, and as a basis for determining peak hour mode share of infill development. (3) Approach is adaptable to meet the needs of emerging multimodal impact analysis techniques. Approaches that require input from regional HTS data or geographic information systems (GIS) data may not yield adequate sample size at the resolution of the peak hour by direction. Data from these sources may be highly limited in categories of land use. May require time/cost to understand database and learn to extract information. Trip cordon counts are compatible, and the most common form of trip data are used. Person-trip cordon counts or site- /building-specific traffic or multimodal counts are compatible and are the most common form of trip data used. Traveler surveys may be required if detailed mode-share data are desired or if site does not fully accommodate its parking demand or if the site has a substantial number of linked trips by drivers who park once and visit multiple sites. Survey costs at complicated sites may be significant. Highly compatible if applied to traditional vehicle or person-trip generation rates or equations. Regression analysis may be limited to a small number of common land use categories. Survey-extracted data may be limited to a small number of common land use categories. The number of records in travel survey database may result in statistically insignificant sample size when extracting data at the finest grain of peak hour by direction. May require time/cost to understand database and learn to extract information. Highly compatible if analyst is collecting data; exact type of data can be collected on an as-needed basis. Minimum required data are percentage non-automobile mode share, average vehicle occupancy, and peak hour person-trip cordon count at building. Requires selection of highly similar site within highly similar context representing project site being studied. Ea se o f u se fo r p ra ct iti on er to a pp ly (4) Input data needed to apply the approach are readily available, or the ease and cost of collecting and applying the data are reasonable. (5) Approach requires no specialized skills or investment in training or software in order to estimate infill trip generation. Uses cordon count or traveler interview data. Correlation of variables and validation may be limited to specific land uses and geographic areas. Independent variable data used in regression analysis usually will come from MPO or other local and reliable GIS databases. Use of rates developed by others with site and context criteria matching project is the least costly and time- consuming method. Developing trip rates using practitioner- collected data; high- cost and time- consuming method. Data collection may include cordon person or multimodal trip counts, site- specific traffic counts, traveler surveys, and independent variable data describing site and context. Traveler interview data or site-specific traffic counts must quantify differences between typical (suburban) and infill development. Correlation of variables and validation may be limited to specific land uses and geographic areas. If adjustment factors representing the appropriate context have already been extracted from travel survey database, then the method is simple to apply. If not, the ease of the extraction process can be moderate to difficult. Requires familiarity with survey database, and being moderately skilled at manipulating large databases, special training, or software may be required. Requires collection of data to develop adjustment factors for baseline ITE data and infill data. Minimum required data are percentage non-automobile mode share, average vehicle occupancy, and peak hour person-trip cordon count at building. Requires general skills in planning, executing, and reviewing data collection efforts and summarizing results.

14 Table 3.1. (Continued). Criteria Grouping Criteria Direct Estimation Based on Regression Analysis Direct Estimation Based on Empirical Data ITE Rate Adjustment Based on Regression Analysis ITE Rate Adjustment Based on Empirical Data Extracted from HTSs Collected at Proxy Sites Cr ed ib ili ty o f a pp ro ac h in te rm s of re lia bi lit y an d co ns is te nc y of p er fo rm an ce (6) Approach applies to the land uses in ITE Trip Generation Manual and has few, if any, restrictions on land use categories and geography. (7) Computations used in the approach are intuitive to the user and transparent by way of documentation. Depends on specific data used in regression model and how well the model has been validated. Use statistically significant sample size for determining desired standard error. Acquiring permission to survey individual sites can add significantly to cost. Selection of sites and intercept surveys may be costly, but once obtained, it is easy to update database. If method is validated, this approach is a credible and reliable estimation method. Method applies to the land uses in the ITE Trip Generation Manual, but may be restricted in terms of land uses and geographic applicability. . Travel survey is a highly credible source of data. Extraction method is limited to a few common land use categories if from an activity-based survey. Traditional origin–destination survey can also be used if the land use at the origin or destination is recorded. Applicable to any ITE LUC. No geographic limitations. Important to ensure that site and context characteristics of selected proxy sites are highly similar and consistent with project site and ultimate context characteristics. A nt ic ip at ed e ffo rt to fu rth er d ev el op o r e xp an d m et ho d to m ee t t he re se ar ch o bje cti ve (8) There is a reasonable cost for collecting data necessary to develop and validate methodology. (9) Approach has the ability to take advantage of existing trip generation database(s) or other readily available data used as input to the method. (10) The principles of development and application of the approach can be described to laypeople. (11) Approach would have likely acceptance by members of the transportation planning and traffic engineering profession who prepare and review site traffic impact analyses. (12) Approach allows for easy addition to or expansion of the database (i.e., new data points or new land uses) once original database is established. Of the four methods, typically requires most data to analyze, along with rate adjustment method using data extracted from surveys. May require significant time to obtain travel surveys from MPOs (typically done every 10 years) or cost to researchers to conduct surveys explicitly to expand method. Ease of explanation varies with complexity of the correlation of independent variables; multivariate correlations more difficult. Principles are readily understood by technical practitioners but are less understandable to policy-oriented reviewers. Cost most associated with data needed for independent variables. Negligible cost for actual application of tool once independent variables are available. Easiest method to understand and explain to the layperson, decision maker, and policy- oriented reviewer. Low-cost, easily applied solution for estimating infill trip generation, but collecting empirical data at sites can be challenging and costly. Cost of independent variable data usually very low; available from developer or property owner. Negligible cost for actual application of tool once independent variables are available. Can use or add to existing trip generation databases if generated using same methodology. Highly credible method; mirrors current ITE trip generation database Analysis cost for regression analysis may be similar for any method based on regression, but depending on the number of independent variables and the extent of validation. Use ITE recommendation of collecting data for at least three or five sites per LUC. Use statistically significant sample size for determining desired standard error. In general, regression analysis is moderately difficult for the layperson to grasp. Regression methods have high likelihood of being perceived as credible by profession. Data extraction from surveys uses existing databases usually available to agencies and consultants. Once data have been extracted for a specific land use within a specific regional context, this need not be repeated if the analyst releases the findings for general use. Explaining the concept of extracting mode share from travel surveys is relatively straightforward, but the actual application of the procedure can be challenging. There is a cost associated with collecting data at proxy sites, but selecting to collect the minimum required data keeps the cost reasonable. Method is easiest of all four methods to explain, and its computational procedure is simple and transparent to reviewers. Because the method starts with an existing credible source of trip generation data, and since the practitioner selects proxy sites and oversees data collection, this method has high probability of professional acceptance. It is not necessary to add to any database; however, practitioners will likely maintain libraries of well- documented proxy site data for use on future impact analyses or for others to use if site meets their criteria. Notes: Examples of independent variables used in a regression analysis are development units by land use, population or employment within development, development characteristics, income levels, vehicle ownership, parking spaces, transit availability, and service.        

15 that develops adjustment factors from the empirical data contained in activity-based household travel surveys (HTSs). Although this second method does not fully meet the research objective or meet all of the selection criteria, it does address gaps in the application of the first method regarding certain types of transportation studies and scale beyond a single site. Thus, the two methods are intended to be employed under different circumstances. In particular, the second method is primarily intended to address: • Adjustment factors for broad classifications of context for use in large-scale impact analyses (region-wide, citywide, or of several hundred acres or more); • Pre-established adjustment factors in guidelines for pre- paring impact analyses for a specific geographic region for consistency and so that practitioners are not required to extract data from HTS databases; and • When there are no sites similar to the proposed infill devel- opment within the same context as the proposed project from which empirical data may reasonably be collected. 3.3 Summary and Conclusion Chapter 3 has documented the process that the research team employed to select the approach and specific methods to develop as part of this research study. The selection process and direction received from the research panel resulted in two methods employing the approach category of ITE rate adjust- ment based on empirical data. 1. Proxy site method – Based on the collection of data at site(s) with similar characteristics and located in similar contexts as the infill development site (the project being studied). The research team developed two variations of this method: (a) The minimum data collection variant outlines an expedited procedure that collects from a proxy site or sites the minimum required data to develop adjust- ment factors. (b) The comprehensive data collection variant is used when a proxy site is situated such that the collection of the minimum required data is not feasible, or when more detailed information about the proxy site’s travel characteristics is desired. 2. Household travel survey method – Derives adjustment fac- tors from empirical data found in the database of a regional HTS. Rather than using data from a specific proxy site or sites in locations representing a project’s context, the pro- cess in this method extracts data representing the desired context from an area at the scale of the traffic analysis zone (TAZ). Data may be extracted from a single TAZ, multiple TAZs, or all of the region’s TAZs representing the desired context. Extraction of data representing specific land uses is based on the activities and trip purposes recorded by the travelers during the survey. A summary of the proposed methods is presented in Table 3.2. Chapter 4 describes the development of the methods and provides guidelines for their application in deriving adjustment factors for base- line ITE trip generation data.

Table 3.2. Summary of methodological approach and application criteria. Method Title and Summary Description When to Use M e t h o d # 1 Proxy site method (minimum data collection) for infill adjustment of ITE trip generation data Adjustment factors derived from the essential minimum empirical data collected through observation and simple cordon counts of proxy sites, and within contexts, similar to the proposed infill development (project). Uses a process to adjust ITE peak hour trip generation rates or trips calculated from ITE vehicle-trip rates. Adjustment factors (mode share and average vehicle occupancy) are derived from surveys at existing sites with the same land uses as the project being studied and located within a context with similar characteristics to the existing or future context(s) in which the project will be located. The objective of collecting the essential minimum data is to collect the least amount of site data that produces reasonable estimates of the percentage of non-automobile mode of travel to/from the proxy site—usually counts conducted at site driveways and building entries. This minimum level of data collection allows a practitioner to use this methodology more quickly, less expensively, and at more sites than the other methods identified in this report. Use this method when: 1. Developing adjustment factors for baseline ITE vehicle-trip generation data as part of the process to convert vehicle-trip data to person-trip data. 2. Estimating peak hour trip generation for infill development as part of the site or transportation impact analysis process when proxy sites are similar to the proposed project and its surrounding context. 3. The location, orientation, site layout, and characteristics of the proxy sites are conducive for collecting the required minimum data. Proxy site method (comprehensive data collection) for infill adjustment of ITE trip generation data Adjustment factors derived from empirical traveler data collected at proxy sites in contexts similar to the proposed infill development (project) using a variety of survey instruments ranging from cordon counts to intercept surveys. Uses a process to adjust ITE peak hour trip generation rates or trips calculated from ITE vehicle-trip rates. Similar to the minimum data collection method, the adjustment factors (mode share and average vehicle occupancy) in the comprehensive data collection method are derived from surveys conducted at sites with similar characteristics to the proposed project except that this method can obtain more detailed information, such as mode share, trip purpose, pass-by trips, trip length, parking location and cost, traveler preferences, and traveler demographics, to name a few common types of data. The survey instruments used to collect traveler data include conventional vehicle counts at site driveways, person-trip counts at building entries, random sample intercept surveys of travelers using the site, mail-in surveys, and surveys using other innovative techniques. Use this method when: 1. A detailed breakdown of mode share other than by automobile/non-automobile is desired, or the practitioner desires traveler data that cannot be obtained from counts. 2. The proxy site or sites do not have exclusive parking facilities or are located where there are nearby but unobservable public or private parking structures, below-ground garages, and street parking. 3. The proxy site or sites experience a high level of linked trips, where travelers who drive park once and walk to multiple sites, and if the practitioner desires to determine the site’s demand for primary versus secondary linked trips. 4. The proxy site has a nearby but unobservable rail station or transit hub, and transit mode share is desired. M e t h o d # 2 Household travel survey method for infill adjustment of ITE trip generation data Adjustment factors extracted from the linked- trip database of a regional HTS. Data are extracted at the geographic scale of the TAZ and filtered by attributes representative of varying contexts. Uses a process to adjust ITE peak hour vehicle-trip generation rates (or direct adjustment of trips). Adjustment factors (mode share and average vehicle occupancy) are derived by extracting certain data from the linked-trip database resulting from a regional HTS. The data are extracted from groupings of TAZs with similar characteristics to the existing or future land use, site, and context(s) in which the proposed project will be located. Extraction filters data by attributes representative of context, land use, mode of travel, time of day, and direction. Mode share and vehicle occupancy data extracted from HTSs can reflect a specific but limited number of land use categories. This method requires having or developing a familiarity with travel surveys, GIS systems, and database manipulation in order to cost-effectively extract the necessary data. In addition, to access the linked-trip database, it is necessary to obtain detailed documentation of the survey’s structure and especially the databases’ library of variables and their definitions. Use this method when: 1. 2. 3. Developing adjustment factors for broad classifications of context for use in large-scale impact analyses (region-wide, citywide, or large districts). Preparing a lookup table of adjustment factors for different context categories to be included in guidelines for preparing impact analyses for a specific geographic region. There are no sites similar to the proposed project and within the same context as the proposed project from which empirical data may reasonably be collected.