Berkeley-Charleston-Dorchester Council of Governments (BCDCOG) Peer Review

5.0 Peer Review Panel Recommendations

This section summarizes the recommendations of the panel.

5.1 Model Assessment

Throughout the morning presentation, the panel was struck by several significant issues with the current BCDCOG model. These are all detailed in Section 4.0 and are listed here for completeness:

• Land use projections: Employment forecasts should be sensitive to presence of retirees and their greater demands for certain employment sectors such as medical, retail and restaurants. Also, land use should be sensitive to the transportation accessibilities.
• Employment forecasts: Employment in the region should not be assumed to be strictly correlated to population, but rather should follow other economic indicators, and the lack of earning level per worker or household income representation in the model should be remedied. Moreover, the employment forecasts should be made through multiple scenarios and not as a single value.
• Retiree households: The growth of retirees is one of the most substantial changes expected in the region, and the model should distinguish these households from other zero-worker households because they have different trip rates, travel patterns, modes, and trip times of day. Due to Charleston's characteristics, retiree households need to be understood in depth.
• Auto-availability: It is insufficient to only use the existence of an auto in the household; instead vehicle availability measures should be used to actually represent captive transit users.
• Restrictive trip purposes: The non-home-based trip purpose should be divided into two purposes: work and other. The home-based-other purpose should be divided into three purposes: home-based-recreation, shop, and other.
• Special Generators: Besides the military base, panel members recommends that BCDCOG remove all special generators and incorporate those trip types into the model.
• Visitor model: Without a visitor model the region is missing a large number of trips based on the size of the tourist industry. The panel agreed that the implementation done in Knoxville to include travel to the Great Smoky Mountain National Park is a good model for BCDCOG as it includes controls based on occupancy level to test demand at different times of the season.
• Distribution model: Panel members agreed with BCDCOG that it is not desirable to use K-factors to capture the deterrence for trips to use leave the peninsula. Panel members believe that the issues with trip distributions, including crossing bridges, can be addressed with a destination choice model and that this can be solved with a relatively simple model enhancement.
• Freight: BCDCOG needs to do more to account for port activity and urban goods movements. Moreover, BCDCOG needs to understand the need for each model component, particularly freight, by engaging model users and regional economy drivers.
• Non-Motorized travel: The share of non-motorized travel in the region is too large to ignore. Also, these trips are included in trip generation rates, and so the trips need to be explicitly accounted for in the model. The consideration of non-motorized trips could be implemented through a non-motorized mode split applied prior to mode choice or by extending the mode choice model to include non-motorized alternatives.
• Employ available data sets: BCDCOG should leverage existing data from SCDOT to do freight analysis and add speed validation to the existing model, but note that doing so would require additional in-house resources to process the data.
• Validation data: The current validation process is inadequate and needs to pay more attention to each component in the model process and not only the end assignment results. Also, mode share and speed should be included in the validation metrics.

These issues led the panel to conclude unanimously that the current model does not meet the state of the practice and that it is not a sound structure from which to build a state of the practice model. Before BCDCOG should attempt to implement the types of improvements they have specified or other sophisticated tools, the model fundamentals need to be structurally sound. Therefore, the panel recommends that BCDCOG continue to maintain their current model to support LRTP and other active planning requests, but that all new development be directed toward building an entirely new model structure. A panel member concluded that the model is not ready for Travel Demand Management (TDM) type analysis nor the upcoming performance-based analysis required by MAP-21.

Panel members encouraged BCDCOG to leverage what they can from existing models. This was reinforced by a panel member who cautioned BCDCOG against building a highly specialized model that is not well founded or justified by the state of the practice. This could leave the agency in an indefensible situation if the model results are challenged.

If a wholesale model development approach is not feasible, BCDCOG could do an incremental approach by reworking each step of the current model. Improvements to the model should be prioritized by the degree of improvement, the resources required, and the time frame. The destination choice model is an example model enhancement that would provide great improvement with a lower level of effort and within a shorter time frame. However, similar to a house reconstruction, it will be significantly more expensive, complex, and time-consuming to improve the model in an incremental way as opposed to starting with a clean slate.

5.2 Model Vision

The panel noted that BCDCOG lacks a clear vision for the CHATS model, and thus it is difficult to formulate a set of guidelines for what the model should include. The panel acknowledges that the model vision should be spearheaded by SCDOT but that leadership at this level could be stronger. The panel recommends that a master model vision be developed to guide future model development. The annual modeling budget should be more specifically allocated and BCDCOG could have their consultants present a menu of items so that they can plan the development process.

The panel was excited by the opportunities present in an area such as BCDCOG, which is a vibrant urban area with large projected growth and active economy. With the proper fundamentals, the panel envisions that this planning group could eventually build a Dynamic Traffic Assignment (DTA) model. The panel concluded that a lifestyle trip-based model would be sufficient for this region and that a tour-based model is not necessary.

5.3 Model Support

Having a state of the practice model requires more than a particular structure and specific model parameters. It also has to do with the perceived usefulness of the model, adoption of the model as a planning tool, and users interested in the outputs. The lack of an active model user group demonstrated to the panel that this model is not well accepted. Several panel members called for greater outreach to decision-makers and other potential model users to understand what they would value in the model.

The panel recognized that there are limited resources even at the state level and recommends that the BCDCOG look to other states for examples and guidance. A panel member encouraged BCDCOG to meet with the Florida modeling team to understand the Florida model process as a representation of how they could operate.

The panel emphasized that BCDCOG and their consultants had done an extraordinary job at maintaining a model with limited resources and support. The panel strongly recommends that BCDCOG increase the internal modeling team substantially in order to build and maintain a state of the practice model. At least one full time senior modeler with substantial prior experience should be hired. This modeler or preferably a second hire would also require strong analytical skills to be able to consume the new data sources being made available.