Highways Agency - Safe roads, Reliable journeys, Informed travellers

Related pages

• 1. Introduction
• 2. Existing Conditions
• 3. Scheme Proposals
• 4. Engineering Assessment
• 5. Environmental Assessment - 1
• 5. Environmental Assessment - 2
• 5. Environmental Assessment - 3
• 5. Environmental Assessment -4
• 5. Environmental Assessment - 5
• 5. Environmental Assessment - 6
• 5. Environmental Assessment -7
• 5. Environmental Assessment -8
• 5. Environmental Assessment -9
• 5. Environmental Assessment -10
• 5. Environmental Assessment -11
• 5. Environmental Assessment -12
• 5. Environmental Assessment -13
• 5. Environmental Assessment -14
• 5. Environmental Assessment -15
• 5. Environmental Assessment -16
• 5. Environmental Assessment -17
• 5. Environmental Assessment -18
• 5. Environmental Assessment -19
• 6. Traffic and Economic Assessment

6. Traffic and Economic Assessment

6. Traffic and Economic Assessment

6.1 Background

To facilitate the identification and evaluation of the preferred route a transportmodel was developed for the detailed assessment of the performance of the A14highway improvement scheme. The CHUMMS based TPI model has been adoptedas a basis for the development of this modelling framework, which is discussed inmore detail in the following sections.

It was noted that the CHUMMS transport strategy also included a scheme for aguided bus system which would make use of the existing A14 west of Fenstantonand the alignment of the dismantled railway line between St Ives and Cambridge.Therefore, the A14 study ensured close liaison with the guided bus study team, andhence consistency in approaches.

This chapter makes regular reference to a number of key locations throughout thestudy area. These are shown in Figure 6.1 in Appendix F.

6.2 Development and Validation of the A14 Transport Modelling Framework

6.2.1 Overview of the Transport Modelling Framework

The transport modelling framework assumed for this study has been describedin detail in both the Work Programme Report⁵ and the Local Model Validation Report⁶ (LMVR). The following sections provide a summary of thedevelopment of the modelling framework used for the assessment of the alternativeA14 improvement options.

Over recent years three similar multi-modal transport models have been developedand used to forecast transport conditions in the Cambridge/Huntingdon area. Theyinclude:

• The County Model, maintained by Cambridgeshire County Council (CCC);
• The CHUMMS Model developed for the purposes of the CHUMMS multi-modal study; and
• The CHUMMS-TPI model, which was based on the CHUMMS model and developed further for the evaluation of the A14 scheme for TPI entry.

The specification of the modelling process adopted for this study was governed bythe level of detail required for the economic and environmental assessments andthe scheme design. On this basis the transport model needed to be capable ofproviding:

• Reliable strategic flows for the identification of scheme options (i.e. for the Technical Appraisal Report (TAR));
• Traffic flows for assessing junction performance and assisting in outline design of the principal elements of the scheme (particular importance in Scheme Assessment Report (SAR));
• Information on changes in performance of the wider area and local highway network for assessments on 'efficiency' (i.e. operational and economic assessments); and
• Traffic flow data for environmental assessments, namely noise and air quality.

A two tier modelling approach was adopted, whereby the CHUMMS model wasused to form the basis of a strategic transport model (STM) for the A14 study area.The purpose of the STM under this study is to forecast future changes in highwaytravel demand for different improvement options and years (by comparing outputmatrices), whilst taking into account the influences of other transport schemes andmajor land-use developments.

The A14 STM is largely an updated version of the CHUMMS model. Whilst theSTM was capable of producing reliable forecast flows for the strategic highwaynetwork, it was not sufficient for representing more local traffic conditions, andhence the performance of smaller junctions along the trunk road.

Therefore, an A14 local highway model has been developed to represent localtraffic conditions within the study area, and hence provide a basis to examine theperformance of the improvement options in greater detail than that offered by theSTM. The highway demand forecasts produced by the STM were fed through toinform the local highway model forecasting process.

The models are designed to output hourly flows for the AM peak (08.00 to 09.00)and for the inter-peak (14.00 to 15.00) for an average weekday.

The development of both the STM and the local A14 highway model is described indetail in the LMVR.

Table 6.1 below summarises the overall structure of the transport modellingframework and how the models link in with other study processes.

6.2.2 Data Collection

Best use was made of a number of sources of transport related data during thecourse of this study. Bearing in mind that the intention was to adapt an existingmodel for the purposes of this study much of the data has been used to check andif necessary enhance the model. In addition to existing datasets, it has beennecessary to undertake surveys to satisfy the data requirements for the modelupdate and enhancements. The collation of existing and new traffic data for theA14 have been reported in full in the earlier phases of this study ^{7 & 8}.

Table 6.1 - Structure of A14 Transport Modelling Process

EXISTING DATA

Recent origin-destination data was available from LATS roadside interview surveysat 15 locations undertaken in 2001 and 2002. Both the interview and traffic countdata associated with these surveys was obtained for use in this study.

Cambridgeshire County Council collect large amounts of traffic count data in thestudy area for their own traffic monitoring purposes. They also monitor generaltraffic patterns over a series of screenlines, some of which have been obtained foruse in the model calibration / validation exercise. The data was collected during2002 and 2003.

Other ad-hoc traffic count data was also available from the County, DfT and othersources. The location of this count data is shown on Figure 6.3 (in Appendix F).This data included the HA TRADS monitoring site at Swavesey on the A14, whichcontains a long term automatic traffic count (ATC) site.

NEW DATA COLLECTION

New link count data was collected to enable two new screenlines to be created.One screenline is just north of the A14 along the disused railway line betweenCambridge and Huntingdon (Figure 6.2 in Appendix F). This will enable the modelto provide a better representation of north-south trip movements to the north of theA14. The second screenline is a Southern Screenline, running close to thesouthern edge of the study area. This enables the model to provide a better controlof trips crossing the southern boundary of the study area.

New traffic count data was also collected on all the A14 slip roads betweenEllington and Fen Ditton, and at the Caxton Gibbet junction between the A428 andthe A1198. This involved data collection at some 30 sites. The collection of newcount data on the A14 slip roads also enabled flow data to be estimated on themain carriageway in addition to the slip roads. (Figure 6.3 in Appendix F shows thelocations of both the ad-hoc surveys discussed above, and the new A14 counts).

New journey time data was also collected along the A14, A428 and the A1198 inboth directions of travel across all modelled time periods. The surveyed routes areshown in Figure 6.4 in Appendix F.

6.2.3 The Strategic Transport Model (STM)

No significant changes to the modelling structure would be made of the CHUMMStraffic model. The original CHUMMS multi-modal model and associateddocumentation was reviewed and the operation of the various components audited.Some minor amendments were required to the model processes, particularly withrespect to the handling of matrices within the mode choice model, but these werefound to have very minor impact on the performance of the mode choice and theresults of the model validation. These are documented fully in the LMVR.

6.2.4 A14 Local Highway Model Development

The development of the local model was required to enable detailed link andturning movement forecasts to be made. It uses the enhanced CHUMMS modelhighway network and matrices. The LMVR sets out in detail the approach adoptedfor the development of the local highway model. A brief summary is given below.

ENHANCE/REFINE HIGHWAY NETWORK AND DEMAND DATA

The CHUMMS highway network has been refined to ensure local movements,especially to and from the local villages and urban centres of Cambridge andHuntingdon, were adequately represented. This mainly involved extending thesimulated area in the modelled network; that is the area in which junctions arecoded in detail so queues and delays are calculated with greater accuracy. Thiswas achieved through a combination of processes, including: detailed coding of keyjunctions within the corridor (based on network inventory data collected as part ofthis study); the inclusion of appropriate speed-flow curves where necessary; and areview of zone connectors to ensure realistic loading of trips on to the network.

Traffic demand was obtained from a combination of the CHUMMS matrices and theLATS roadside interview surveys. The final matrix details were enhanced usingmatrix estimation techniques.

ASSIGNMENT PARAMETERS

The model assignments have been carried out using a Wardrop User Equilibriumprocedure, which seeks to minimise travel costs for all vehicles in the network. Theassignment is based on minimum generalised cost routes where the generalisedcost is defined as a linear combination of time and distance:

Generalised cost = time + a x distance

The coefficient of travel time is set by default to 1.0. Based on advice given inHighways Economics Note 2 (HEN2) - DMRB Volume 13, the distance element(? ?of generalised cost functions for trip routeing in the AM peak period was taken as0.25 for cars and light goods vehicles, and 0.72 for heavy goods vehicles. In theinter-peak period the values were 0.28 and 0.77 respectively (See the LMVR forfurther information).

Based on Guidance set out in DMRB the calibrated route choice criteria used in thebase year model have been assumed to remain unchanged in the future.

There is a point of view that some drivers base their route choice solely onminimising time. A test has been undertaken with the distance parameter set tozero, which assumes that driver route choice is based entirely on minimum time.The test showed no significant change to the assignment results, with the numberof links passing both the GEH and flow criteria tests remaining virtually constant(See LMVR). The conclusion of the test was that the base model was insensitiveto the distance parameter, and that both the HEN2 value of distance and a value ofzero were capable of satisfying the DMRB criteria. However, to be consistent withDMRB guidelines, it was decided to use the HEN2 values of distance as a basis ofthe routeing parameters for the model.

However, it was thought that this could nevertheless be an issue when modellingthe A14 improvement options, given the increased route choice betweenCambridge and the M11 in the east and the A1 and Huntingdon to the west. Theabove tests were repeated assuming a Do-Something scenario, but again it wasfound that the model (and hence the routeing) was not sensitive toinclusion/exclusion of the distance coefficient.

MODEL CALIBRATION AND VALIDATION

The main aim of this task was to match modelled flow patterns and travel times inthe local highway model with recently observed data, and thereby demonstrate thatthe model provides an adequate representation of the current situation and issuitable for forecasting future traffic conditions along the A14 corridor. The resultsof the local model calibration and validation exercise, which have been presentedin full in the LMVR, satisfy the DMRB criteria for the validation of traffic models.

6.3 Traffic Forecasting and Option Testing

6.3.1 Overview of Forecasting Process

A number of forecasting procedures have been developed to predict future growthfor the following components of the demand matrices:

• Light vehicle trips relating to the study area (internal-internal; internal-external and external-internal trips) based on CHUMMS compatible processes;
• External to external light vehicle trips - based on NRTF car forecasts, but adjusted to reflect the higher expected level of growth in the region relative to national expectations; and
• Heavy vehicle trips (all trips) - Assumes NRTF predictions for light goods vehicles (LGV) and other goods vehicles (OGV).

The STM combines the functions of the CHUMMS/MENTOR land-use model (tripgeneration and distribution) and the highway/public transport mode choice andassignments models. For each model run the generalised costs of travel for allorigin-destination movements, as extracted from the highway and public transportassignments, are fed back in to the land-use model. The land-use model then reevaluatesthe trip generation and distribution based on these updated travel costs,and hence accessibility to and from different zones.

However, the MENTOR land-use transport interaction model is only capable ofproducing forecasts for study area light vehicle trips. Future changes in externalexternallight vehicle demand, and all commercial vehicle trips were forecastoutside the MENTOR procedures, and were based on guidance relating to NRTFand TEMPRO forecasts.

The use of the MENTOR model to produce travel demand forecasts by mode,ensured that the assumptions inherited through the CHUMMS model wereretained. However it was necessary to provide updated information relating to anumber of key parameters. These are covered in detail in the Forecasting Report9,and comprise:

• Land-use assumptions - A key change that has been made under the A14 improvement study involved updating the land-use targets (households and employment) to reflect the County's Structure Plan Review (October 2003);
• Travel costs between all origin-destination pairs, reflecting proposed transport improvement schemes; and
• A number of parameters reflecting predictions on economic activity in the future.

One land-use development that is of significant importance to the A14 improvementscheme is that proposed at Northstowe. This is a new settlement sited some 3 kmto the north east of Bar Hill. The development will be phased in over a number ofyears with plans indicating that in the region of 10,000 dwellings could be providedin this area by 2025. The County Structure Plan Review shows a provision of 6,000additional dwellings by 2016. It is anticipated that the developers promotingNorthstowe settlement will be submitting a planning application for up to 8,000dwellings which are expected to be in place by 2021. For the purpose of this studythe forecasting assumes that aspirations for 10,000 dwellings in this area will berealised (that is a further 2,000 units between 2021 and 2026).

GROWTH ASSUMPTION BEYOND 2016

As already stated, MENTOR relies on input from the Structure Plan which sets outplanning policy up to 2016, and hence is only capable of producing forecast traveldemands based on land-use patterns in the Cambridge Sub-Region up to 2016.

There is a close agreement between the growth factors derived through MENTORfor the study area (2003 to 2016) and the TEMPRO policy based predictions forCambridgeshire. Beyond 2016 it was therefore considered most appropriate tobase future traffic predictions on TEMPRO (Version 4.2.3) policy based trafficforecasts, adjusted by the relative difference between MENTOR and TEMPROover the period 2003 to 2016. These growth factors were then applied to all studyarea trips (i.e. internal to internal, internal to external and external to internal).

It is important to note that the element of the proposed Northstowe developmentexpected to take place beyond 2016 was modelled explicitly in the model. The post2016 development of Northstowe was represented in the model on a pro-ratabasis, that is by factoring the 2016 trips associated with Northstowe in line with theadditional housing between 2016 and 2025. The TEMPRO based factors(discussed above) were adjusted to net out the growth attributed to the additionalNorthstowe development before being applied across the remainder of the studyarea.

6.3.2 Years of Assessment

For the purposes of evaluating these improvement options, forecasts wererequired, from a base year of 2003, for an opening year of 2010 and a design yearof 2025. An intermediate year of 2016 was also forecast.

6.3.3 Forecasting Requirements

For the purposes of developing and evaluating the improvement options it wasimportant that the forecasting models were capable of generating the trafficinformation required by the different study processes. These include the following:

• Economic evaluation;
• Environmental assessments;
• Scheme designs;
• Landscape; and
• Drainage assessments.

The processes above called for a combination of 18 hour and 24 hour AADT and/orAAWT traffic flow forecasts. Therefore factors were applied to expand the hourlymodelled flows (AM peak and inter-peak) to 24 hour AADT and 18 AAWT flows.Details on the derivation of the 24 hour AADT and 18 hour AAWT expansionfactors are presented in the Forecasting Report.

6.3.4 Do-Minimum Forecasts

Major transport schemes must be evaluated against an appropriate Do-Minimumbase case. The Do-Minimum scenario has been developed to represent the mostrealistic view of future transport conditions without any new A14 transportproposals. In this scenario only committed transport schemes and land-useproposals that are likely to be completed by each forecast year have beenconsidered. In addition any schemes that would be required should the A14improvement scheme not be implemented have also been identified and included.Only those proposals which are likely to have a material effect on travel patternsand/or mode share within the A14 corridor have been represented directly withinthe model. The strategic effects of some smaller measures, such as local trafficmanagement schemes, are likely to be negligible and hence not included. Theassumptions underlying the Do-Minimum forecasts are fully documented in theForecasting Report.

Huntingdon Viaduct

In addition to committed transport improvements the Do-Minimum is also based onthe premise that the Huntingdon Viaduct will be replaced with a similar structure(i.e. dual 2 lane carriageway) should the A14 scheme not be built.

Northstowe

The proposed Northstowe settlement itself is not actually a committed developmentitself. However, based on housing allocations set out in the County Structure Plan ithas been assumed as committed for the purposes of this study.

To cater for development related traffic the Do-Minimum assumes the upgrading ofthe B1050 carriageway to a dual 2 lane standard between Northstowe and theexisting A14, feeding into a new junction at Bar Hill. The highway network in thevicinity of Bar Hill will also be improved with the provision of a local access roadrunning parallel to the A14 on its northern side, connecting the existing A14junctions at Bar Hill and Dry Drayton.

Forecast Do-Minimum Traffic Growth and Flows

By 2010 under the Do-Minimum scenario, (i.e. assuming the A14 improvementscheme is not built), highway demand across the study area is forecast to rise by8% to 9% compared to the 2003 Base Year. These estimates are based on the"most likely" or central growth forecasts. Similarly, by 2016 and 2025 traffic levelsare forecast to increase by over 15% and over 25% respectively. These figuresrelate only to study area related trips and exclude external-external trips.

Forecast traffic flows (AM, inter-peak hour and 24 hour AADT) for different sectionsof the primary highway network serving the study area are shown in Figures 6.5and 6.6 (Appendix F) for years 2010 (opening year) and 2025 (15 years afteropening) respectively. All flows shown relate to a central growth scenario. Forcomparative reasons 2003 base year flows are also provided (Figure 6.7). Table6.2 provides a summary of expected levels of daily traffic growth (24 hour AADT)for key sections of the A14 trunk road, where the growth factors can be comparedwith the predicted increases in overall highway demand of over 25% by 2025(central growth).

Clearly, capacity limitations of the A14 between Godmanchester and Bar Hill wouldconstrain traffic growth in this area, despite more significant growth occurring onupstream and downstream sections of the trunk road network. Typically, betweenGodmanchester and Bar Hill daily traffic growth of around 10% can be expected by2025, with Do-Minimum assumptions,. On the other hand, higher levels of growth indaily traffic flows can be expected by 2025 for the A14 between GodmanchesterJunction and Brampton Hut (around 26%) and on the Cambridge Northern Bypass(CNB) (28% to 30%).

The AM peak forecasts indicate an increase in traffic demand of over 25% by 2025(central growth). As a result, motorists in the A14 study area would experiencedeterioration in travel conditions during peak travel times (Table 6.3 in Appendix F),for example:

• Across the highway network over-capacity queuing would increase eleven-fold;
• Transient queuing and delays would rise by 75%;
• Total travel time across all journeys in the study area would increase by over 76%; and
• Average travel speeds across the highway network would reduce by 29% to 43 kph in 2025, compared with 60 kph in the base year (2003).

FORECAST DO-MINIMUM JOURNEY TIMES

Due to the relatively modest growth in traffic on the A14 by 2010 the effect onaverage journey times along the 38 kilometre section of road will be limited to anincrease of 3 minutes (10%) in the westbound direction and 9 minutes (25%)eastbound (Figures 6.8 and 6.9 in Appendix F).

By 2025, journey times over the full length of the scheme are expected to increaseby 11 minutes in the westbound direction and 22 minutes in the eastbound, a riseof 33% and 67% relative to the base year. In the morning peak direction(eastbound) travel speeds are affected mostly between Godmanchester and BarHill.

6.3.5 Do-Something Forecasts

The forecasts produced for each of the A14 improvement options show that travelconditions can be improved through the introduction of a suitable scheme. Theforecast flows are summarised in Table 6.2 and shown in more detail in Figures6.10 to 6.17.

Across the study area highway network the 2025 central growth forecasts for AMpeak hour (Refer to Tables 6.4 to 6.6 in Appendix F) show that, all the optionsconsidered would:

• Deliver a reduction in over-capacity queuing, relative to the Do-Minimum, of over 16%, though there would still be a eight-fold increase over the base year;
• Improve transient queuing by between 5% and 9% compared to that forecast for the Do Minimum;
• Reduce total travel time by around 6%, and hence increase the average speed to around 49 kph from 43kph in the Do Minimum (+15%).

But the most prominent benefit is the improvement in journey times offered by eachoption along the A14 between Ellington and Fen Ditton. With any of the schemes inplace journey times not only improve relative to the Do-Minimum, but forecastsindicate time savings of 6 minutes (-20%) over the base year in the westbounddirection during the morning peak in 2025, and around 9 minutes (-25%) eastbound(Refer to Figure 6.18 and 6.19 in Appendix F). These savings are largely generatedby the improvements offered by the scheme in the AM peak between Fen Dittonand Girton in the westbound direction (AM), and Godmanchester and Bar Hill in theeastbound direction. The figures also illustrate that the A14 schemes would operatewith uniform speeds along its length, thereby implying improved journey timereliability.

Table 6.2 - Summary of Forecast Daily Flows (24 Hour AADT): 2025 (Central Growth) Single Direction (20KB PDF)

6.3.6 Scheme Operational Performance Summary

The operational performance of each scheme is very similar, with little between theforecast journey times along the entire routes. There is however a number ofrelatively minor differences which, when accumulated over the whole scheme, doenable meaningful comparisons to be made between the various schemes.

In assessing the preferred scheme comparisons there are the two different routes,i.e. Orange and Purple, and then within each route option there is a Full Junctionoption and a Limited Junction option.

ORANGE ROUTE COMPARED WITH THE PURPLE ROUTE

The Orange Route performs better than the Purple Route for three basic reasons:

• To accommodate high turning movements between the improved A14 and the old A14, the free-flow interchange at Fen Drayton in the Orange Route works better, with lower levels of delay than the Galley Hill Junction in the Purple Route, which is controlled by a conventional high capacity roundabout;
• In the Purple Route, the A14 corridor between Galley Hill and Trinity Foot has less capacity than the Orange Route highway layout. Consequently, in the Purple Route, the network in this area is operating closer to capacity, with resulting slower speeds, which reduces the attractiveness of this section of the network; and
• The Orange Route is slightly shorter than the Purple Route.

FULL JUNCTION OPTION COMPARED WITH THE LIMITED JUNCTION OPTION

The Limited Junction Option performs better than the Full Junction Option for threebasic reasons:

• Although there is no direct connection between Bar Hill and the improved A14, the local access road configuration adequately caters for Bar Hill and Northstowe traffic;
• The local access road in the Limited Junction Option has grade separated layouts at Bar Hill and Dry Drayton. This removes the delays that occur at both these locations in the Full Junction Option; and
• In the Limited Junction Option the local access road has better connections into the rest of the highway network at Girton, including direct links to the A14, M11 and Huntingdon Road. In the Full Junction Option the local access road only connects directly with Huntingdon Road.

6.3.7 Required Highway Provision

The Highways Agency's guidelines to determine carriageway capacity are basedon AADT flows for the assumed opening year. In this case forecasts relating to a2010 central growth scenario have been adopted.

Traffic flow ranges for use in the assessment of new rural roads are given in theDesign Manual for Roads and Bridges (DMRB) Advice Note TA 46/97. The notegives the economic assessment and recommended flow ranges for new rural roadlinks for opening year AADT flows relating to various carriageway standards. Theseare set out in Table 6.7 as extracted from TA 46/97.

The opening year AADT flows for key sections of each A14 Improvement option,together with the required carriageway standard (as shown above) are given inTable 6.8.

Table 6.8 - Required Carriageway Standards by Option (20KB PDF)

6.4 Overview of Economic Assessment Process

The economic assessments for each improvement option have been carried out inline with Department for Transport (DfT) and Treasury guidance as detailed on theDfT Transport Appraisal website¹⁰ and in the Transport User Benefit Appraisal(TUBA) guidance¹¹.

The wider economic impacts of the improvement (including its regenerationpotential) are not considered in the assessment as they are not considered to besignificant in this area, a view supported by the HA.

The economic evaluation of the A14 improvement schemes has been fullydocumented in the Economic Assessment Report (EAR)¹². This section provides asummary of the approach adopted for the economic assessments, together withthe underlying assumptions and the results.

6.4.1 Costs and Benefits Considered

The economic assessment of each improvement option was based on acomparison in monetary terms of the total benefits generated by the option againsttotal associated costs. The following four elements were covered by thecomparison:

• The impacts of each option on travel times and travel costs for trips made within and through the modelled study area together with the associated impacts on revenue and indirect tax levels. These impacts were estimated on the basis of the forecast change in travel conditions caused by each option compared to a Do-Minimum scenario. Conditions in each scenario were forecast using the A14 transport modelling framework, with outputs from the models used to calculate traveller user benefits, revenue and indirect tax benefits using the DfT's TUBA program;
• The impacts of each option on road accidents in the study area. These were estimated using COBA 11¹³ methodology and changes in traffic levels forecast by the transport models;
• The capital and maintenance costs associated with each option. These were derived on the basis of reference schemes and standard rates. Allowances were also made for risk and optimism bias;
• The impacts of the construction and ongoing maintenance programmes on travel times and delay for journeys within and through the study area; estimated using the local highway model and a single year application of the TUBA model.

6.4.2 Forecast Years and Appraisal Period

For each improvement option, estimates of each element of the scheme's costsand benefits were made for three forecast years:

• 2010, opening year;
• 2016, an intermediate forecasting year; and
• 2025, design year (15 years after opening).

These forecast costs and benefits were then combined within the TUBA program toproduce an overall estimate of the balance of the costs and benefits of each optionover a 60 year appraisal period (2010-2069).

6.4.3 Overview of TUBA Process

TUBA (Transport User Benefit Appraisal) is bespoke software developed on behalfof the Department for Transport (DfT) to estimate the multi-modal impacts oftransport schemes in terms of the costs and benefits experienced by users andproviders of the transport system and the associated indirect taxation and subsidyimpacts. All impacts are considered in monetary terms. For the purposes of thisstudy TUBA version 1.6A has been used.

Details of the processes adopted in TUBA, together with the underlyingassumptions are discussed fully in the EAR. The following paragraphs provide avery broad summary of the way in which the TUBA software was used under thisstudy.

The program estimates costs and benefits experienced by users and providers ofthe transport system by comparing transport conditions in a Do-Something (DS)scenario against conditions in a Do-Minimum (DM). To this end, TUBA usesinformation from the transport models to:

• Calculate user benefits by mode and for each element of journey cost. (i.e. travel time, vehicle operating costs (fuel and non-fuel) and travel charges (fares and parking)); and
• Calculate the change in revenue generated by the transport system and changes in the indirect tax income received by the government (on the basis of different levels of indirect taxation incurred on, for instance, fuel costs and PT fares).

The user and provider related costs and benefits in each year are then combinedwith estimates of capital and maintenance costs and accident savings (calculatedin separate parallel processes described below) and discounted to 2002 values(using a discount rate of 3.5% for the first 30 years and 3.0% thereafter). Thediscounted values of the benefits are then summed to give the Present Value ofBenefits (PVB) and used to calculate the Net Present Value (NPV) of the schemeby subtracting the Present Value of Costs (PVC) of the scheme: i.e. NPV = PVB -PVC.

For each assessment run TUBA produces summary results in the form of theTransport Economic Efficiency (TEE) and Public Accounts tables required by theDfT's revised New Approach to Appraisal (NATA).

INPUT PARAMETERS AND ASSUMPTIONS

The EAR sets out the key assumptions and parameters used for running each ofthe A14 TUBA assessments, along with their sources. Most of the values adoptedwere based on the guidance given in TUBA, although local data was used whereavailable and relevant, for example:

• Purpose splits: Estimates of the proportions of highway journeys undertaken for business, commuting and other purposes were derived from information from the Road Side Interview (RSI) surveys undertaken at 15 locations within the study area in 2001/02 as part of the LATS¹⁴ survey programme;
• Vehicle Occupancies: Estimates of the average number of passengers in each vehicle travelling by purpose and time period were again based on the LATS RSI data;
• Vehicle Type Proportions: Estimates of the relative proportions of cars, light goods vehicles and heavy goods vehicles were derived from a series of more than 20 manual classified counts (recording traffic flows by vehicle type and time period) at locations within the study area during neutral months (April, May, June, September and October) ¹⁵ in 2003;
• Annualisation Factors: These were used to convert estimates of demand related costs and benefits experienced during the two modelled time periods (AM peak hour (8-9 am), and inter-peak hour (2-3 pm)) into estimates of total annual costs and benefits for each modelled year. The factors were based on relative levels of travel demand for each mode at different times of the day and week (i.e. weekday and weekend):
  • For highway modes, this information was derived from comprehensive long term local traffic count information from the seven HA TRADS¹⁶ survey sites located on the A14 within the study area;
  • For public transport modes there was no equivalent, reliable local data available so national average data from the National Travel Survey (NTS)¹⁷ was used;
  • For park & ride the factors were based on demand profiles derived from patronage surveys undertaken at the five Cambridge based park & ride sites in October 2003.

These factors were used for most elements of costs and benefits but alterationswere made for (see Economic Assessment Report for more detail):

• The peak highway time factor: to account for the fact that congestion levels were not consistent throughout the peak period and were likely to be considerably worse in the modelled peak hour than in the shoulder hours. This adjustment was derived through a comparison of the scale of average time savings (between the DM and DS) experienced in (i) the peak hour (8-9 am) and (ii) the modelled pre-peak hour (7-8 am)¹⁸; and
• The inter-peak based parking revenue factor: to account for the fact that parking charges are lower in the evenings than in the middle of the day (i.e. the interpeak modelled hour).

6.4.4 Safety Assessments

TUBA does not calculate accident savings so the impact of each improvementoption on the number of accidents in the study area was estimated separatelyusing the spreadsheet safety model developed under CHUMMS, updated for useon this study. This spreadsheet is based on the COBA11¹⁹ recommendedmethodology for calculating accident numbers and costs (updated to include the2003 revised values and 2004 revised costs). Observed local accident ratesderived from accident records for the five year period to December 2002 were usedfor the existing A14, with default COBA rates by road type used for other roads inthe study area, including the scheme. Default COBA casualty rates by road typeand average costs per accident and casualty were used for all roads.

The COBA rates are in terms of accidents per million vehicle-kilometres andcasualties per accident. They also reflect an assumed general decline in theincidence and severity of accidents, in line with recent trends in and policies forroad safety. The costs through time include an assumed growth in value, in linewith growth in gross domestic product (GDP).

To estimate accident savings, each road in the A14 modelled highway network wascategorised according to the different COBA road types (based on roadclassification, age, speed limit and number of lanes). Using information extractedfrom the local highway models, total vehicle-kilometres travelled by road type werederived and multiplied by the corresponding accident and severity rates for thatroad type to estimate total numbers of accidents and casualties by scenario. Theaverage costs per accident and casualty, as defined in COBA, were then applied toproduce total costs for all accidents by scenario. The accident cost savingassociated with each of the options was estimated by comparing the respectivetotal accident costs with the Do-Minimum costs.

These calculations were repeated for each of the three forecast years and wereconverted into an estimated Net Present Value (NPV) of accident savings over a 60year appraisal period (2010 - 2069) using the same principles as applied in TUBA.

No assessment of public transport related accident savings was made.

6.5 Construction and Maintenance Costs

6.5.1 Capital Costs of Schemes

Detailed estimates of the capital cost of each improvement option have been madeand are shown in 6.10. Each estimate covers the costs of roadworks, landscaping,impacts on utilities, land acquisition, noise compensation, andpreparation/supervision.

The road works cost estimates have been derived from the measurement ofquantities of each item from design information and appropriate unit cost rates fromvarious reference projects and the HA's new works rates database.

The soft landscaping costs were estimated on the basis of experience of similarschemes, whilst the utility cost estimates were provided by the relevant utilitycompanies where available (in accordance with the New Roads and StreetworksAct, 1991). Some provisional assessments were also made where the estimateswere not forthcoming.

The land acquisition cost estimates were based on detailed drawings of land takeand an understanding of the type of land involved.

At this stage, in the absence of more detailed information, the HA Scheme Budgetvalue has been used for likely costs of noise compensation. However the SchemeBudget values for preparation and supervision costs have been revised and theyare now assumed to be 8% and 5% of total works costs respectively.

Note LJO = limited junction option

6.5.2 Risk and Optimism Bias Allowances

Detailed analysis of the financial risk associated with the cost estimates has alsobeen undertaken, primarily through a Value Engineering and Risk ManagementWorkshop for the study held in December 2003 and an ongoing live risk register.The validated figures from the register were fed into a risk model (using 'At Risk'software) which quantified the risk associated with the scheme's costs as £54.1million for non-land costs and £12.3 million for land costs (2001 Q3 prices), at the50% confidence level. This allowance was included in all the cost estimates usedin the economic assessment process.

Guidance on appraisal and evaluation is given in the HM Treasury's Green Book²⁰.Consequently, an allowance was applied to costs to account for optimism bias.Optimism bias is defined as the systematic tendency for project appraisers tounderestimate their scheme's cost and therefore overestimate the strength of itseconomic case. The allowance made for this bias can be reduced as the schemeprogresses and knowledge improves. So, in line with the HA Guidance²¹ on theappropriate optimism bias allowance for Standard Highway Schemes which are atthe Public Consultation stage and where a risk assessment has been undertaken,a figure of 5% was applied to all costs (including risk) for this study.

6.5.3 Do-Minimum Costs

Most of the costs involved in implementing the Do-Minimum scenario would also beincurred in implementing the Do-Something scenarios because several schemeswould appear in both scenarios, for example the Cambridge Guided Bus. Thesecosts can therefore be ignored in the comparison of the Do-Something routeoptions against the Do-Minimum.

However, an estimated £45 million of the costs (2001 Q3 prices, including optimismbias, VAT and inflation allowances)²² relating to the Do-Minimum would not beincurred should the A14 Do-Something scheme be implemented. These savingsare generated from:

• Huntingdon Viaduct Replacement. In the Do-Minimum scenario it is assumed that the replacement for the viaduct would be similar to the existing structure (a dual 2 lane viaduct). However, with the introduction of the A14 improvement scheme a smaller scale replacement structure would be adequate. In the Do-Something scenarios, it is therefore assumed that the viaduct would be replaced with a single 2 lane carriageway together with an atgrade junction with Brampton Road. This smaller scale Do-Something replacement would be less expensive than the Do-Minimum structure.
• Northstowe Access at Bar Hill. The improvement options include access to the highway network for the traffic generated by the proposed Northstowe development. This therefore removes the need for the separate access scheme designed by the developers for a scenario with no improvement in place (including alterations to the Bar Hill slip roads and construction of a new roundabout and local access road). It is likely that the developers will contribute to the cost of these improvements either in the Do-Minimum or a Do-Something scenario.

6.5.4 Maintenance Costs

The improvement options will increase the maintenance costs associated with theroad network through the provision of a new section of road and the addition oflanes to existing roads. Estimates of these extra costs have been made using thedefault assumptions given in COBA11²³. COBA assumes that maintenance costscan be split into the two broad categories of traffic related costs (which areassumed negligible) and non-traffic related costs which are estimated on the basisof default average annual costs per kilometre by road type.

The additional annual maintenance costs relating to the improvement options werederived based on estimates of the length of new carriageway and existingcarriageway subject to widening. The annual costs were then converted into NPVsin 2002 values and entered in the maintenance section of the Public Accountstable.

6.5.5 Impacts of Construction and Maintenance Works on Travel Times

The economic impacts of the construction process are expected to be primarilycaused by the delay and re-routeing due to bridge and road closures and speedrestrictions. These impacts would increase vehicle travel times. Similarly, ongoingmaintenance on the improved A14 and the replacement of the Huntingdon railwayviaduct would also cause disadvantages through re-routeing, reduced averagespeeds and increased travel times. The impacts of these effects have beenestimated using the local highway model and single year runs of the TUBAprogram.

This approach was used in preference to the more conventional approach of usingthe QUADRO program because it provides a more realistic representation of theimpact of construction and maintenance works over the wider study area.QUADRO, with its single fixed diversion routes, could not reflect the range of localtravellers' responses to any road closures and/or restrictions. The local highwaynetwork offers a range of alternative routes for different movements, and hencelocal travellers are likely to use a variety of routes rather than a single diversion.

Details of the assumptions adopted in running the local highway model inconjunction with TUBA, including the representation of the construction andmaintenance works, are presented in full in the EAR.

In summary, for each modelled maintenance or construction situation, highwaydemand, time and distance matrices were extracted from the model and used, inconjunction with equivalent data relating to a scenario without construction ormaintenance works, in a single year TUBA run.

These single year TUBA runs followed exactly the same principles as thosedescribed for the main TUBA assessments outlined above. The only differenceswere that:

• Benefits were only input and estimated for a single year (not 60);
• 'Annualisation' factors were considerably smaller because the single hour benefits from the highway model were only being factored up to represent the short period for which that closure/restriction would apply;
• The public transport modes were not included as the likely impact of the works upon them was assumed to be slight in comparison with the impacts on private vehicles.

The results of all the tests were combined to create a total NPV of the costs causedby construction and maintenance works in each option for inclusion in the finalTransport Economic Efficiency (TEE) and Public Accounts tables and the economicassessment.

6.6 Economic Performance of Options

6.6.1 TEE, Public Accounts and Summary Analysis Tables

Tables 6.11, 6.12 and 6.13 provide the summary information from the TransportEconomic Efficiency (TEE), Public Accounts and Summary Analysis Tablesrespectively for each of the improvement options. The tables summarise the costsand benefits associated with each option during the 60 year period followingopening and are disaggregated by benefit type, purpose and vehicle type. TheTEE table (Table 6.11) summarises the impact of the option on transport users andthe private sector whilst the Public Accounts table (Table 6.12) summarises theimpact on the public sector.

The overall summary results (including the NPV and benefit to cost ratios) areshown in Table 6.13. The NPVs indicate that there is a strong economic case for allsix improvement options. The strongest case is shown for the Orange and BlueRoute options which yield NPVs of around £2.0 billion compared to approximately£1.8 billion for the Purple Route options.

Overall the pattern of costs and benefits are very similar in all options, with themain variations being in terms of scale. It is therefore possible to highlight severalkey issues from the tables which are common to all route options:

• There is a strong economic case for the schemes. Both the BCRs (showing the ratio of benefit to central government's cost) and the BKRs (showing the ratio of benefit to the HA's cost) are considerably greater than 1 showing that the monetised benefits of all options exceed their costs. In all cases the BCR is greater than 14.7 and the BKR greater than 5.4.
• In all options, the majority of benefits are derived from highway time savings. Along the A14 corridor average journey times will reduce by around 20 minutes eastbound in the AM peak (7 to 8 minutes westbound) and 8 to 9 minutes eastbound in the inter-peak (5 to 6 minutes westbound) for those using the new route.
  
  During an average weekday, the value of savings accrued during peak hours are between 10% and 20% greater than those accrued during the non-peak. The annual weekend (and bank holiday) savings equate to approximately 20% of the total value of savings experienced over the working days of the year.
• The options generate accident savings. The implementation of the A14 improvements has two main impacts which have opposing influences on accident numbers. The introduction of a new, high quality road link on one hand will tend to result in fewer accidents, but on the other hand, increasing the average distance travelled for many trips and attracting more traffic to the corridor will act to increase the number of accidents. The combined effect of these impacts is to reduce the number of accidents by approximately 140 personal injury accidents (PIAs) by the year 2025, and 2900 total accidents including PIAs and damage only accidents. Whilst in percentage terms this equates to less than 3% of accidents, in monetary terms these savings represent an NPV of around £300 million, nearly 15% of the overall present value of benefits (PVB) of the schemes.
• There is a net increase in vehicle operating costs. The magnitude of the total increase equates to around 12% of the monetary value of the time savings for each option and the increase is primarily experienced on freight trips (over 90% of the total) which incur considerably higher vehicle operating costs than car and light vehicle trips. Although the absolute increase is fairly large, it only represents a very marginal change in the total vehicle operating costs associated with highway travel in the modelled area (around 1%). The changes are primarily due to the slight increase in average distance travelled per journey, as the new A14 route influences travel patterns causing drivers to make more longer distance trips as they take advantage of the improved route.
• The options generate increased indirect tax revenues for central government. These are largely due to increased expenditure on fuel and nonfuel related highway costs.
• All options cause a slight reduction in public transport revenue. This reflects a slight decrease in public transport patronage as the A14 highway corridor performs better with the scheme in place.
• There are slight park & ride travel time savings and vehicle operating cost losses for each option. These are generated by the car legs of the journeys which experience the same distance changes discussed above for car only journeys.

6.6.2 Reliability

The economic assessment presented does not include a monetised estimate ofreliability benefits because the methodology for quantification is still underdevelopment and any results are considered provisional. However, each of theA14 improvement options should produce reliability improvements through anumber of effects.

Firstly, each option would provide increased capacity and therefore reduced stresson the network, improving performance and providing more flexibility to cope withfluctuations in capacity and demand and with incidents. Additionally, throughproviding a new section of road, the improvements would provide a new alternativeroute for a long section of the corridor, helping to mitigate the impacts of incidentsby providing more options for diversion routes. Finally, the improvements areforecast to reduce accidents and therefore the number of incidents occurring andinfluencing journey time reliability.

• Table 6.11 - Summary of TEE Tables for All Options
• Table 6.12 - Summary of Public Accounts Table for All Options
• Table 6.2 - Summary Analysis Table for All Options
• download all 3 tables (25KB PDF)

5. A14 Improvement Ellington to Fen Ditton, Work Programme Report, by Atkins on behalf of theHighways Agency (August 2003) back [5]
6. A14 Improvement Ellington to Fen Ditton, Local Model Validation Report, by Atkins on behalf of theHighways Agency (April 2003) back [6]
7. A14 Improvement Ellington to Fen Ditton, Initial Traffic and Accident Data Report, by Atkins on behalf of HA (July 2003) back [7]
8. A14 Improvement Ellington to Fen Ditton, Data Collection Report, by Atkins on behalf of HA (March 2003) back [8]
9. A14 Improvements Ellington to Fen Ditton, Forecasting Report (Draft), by Atkins on behalf of theHighways Agency (June 2004) back [9]
10. http://webtag.org.uk back [10]
11. TUBA User Guidance and User Manual, Version 1.6A, prepared by Mott MacDonald on behlf of theDfT, June 2004. TUBA is the DfT's bespoke software for carrying out economic assessments of theimpacts multi-modal transport schemes back [11]
12. A14 Improvement Ellington to Fen Ditton, Economic Assessment Report (Final Draft), by Atkins onbehalf of the Highways Agency (October 2004) back [12]
13. The COBA Manual: DMRB, Volume 13, Section 1, Part 2: The Valuation of Costs and Benefits back [13]
14. The LATS RSI surveys were a series of RSIs undertaken by the HA and LATS (London Area Transport Study) during neutral months in 2001 and 2002 with the intention of capturing strategic trips and trunk road traffic. The surveys occurred over the whole of the South East of England and included the 15 survey sites used which fell within the A14 model study area and comprised a screenline running north of Huntingdon, St Ives and Cambridge and selected interchanges on the A14 and M11. back [14]
15. Neutral months are those in which flows are most stable and the impacts of seasonal factors likely to influence traffic patterns (such as school holidays) are minimised. They are defined as April, May, June, September and October in Volume 13 (COBA Manual) of DMRB; Part 4, Para 6.3 back [15]
16. TRADS: A database of motorway and trunk road count data maintained on behalf of the Highways Agency back [16]
17. National Travel Survey: Focus on Personal Travel 2001, Table 7.14. DfT, 2002 back [17]
18. The pre-peak hour was modelled to provide estimated queues for the start of the peak period. The model was identical to the AM peak but with an assumed 90% (based on local count data) of the peak hour demand. back [18]
19. COBA Manual: DMRB, Volume 13, Section 1, Part 2, Highways Agency 2002 back [19]
20. The Green Book, Appraisal and Evaluation in Central Government, HM Treasury, 2003 back [20]
21. Full PAR3 Guidance, Full Project Appraisal Report, Version 3.2 Guidance Note, Highways Agency,July 2003, Appendix C, Table C1 back [21]
22. It is noted that these cost estimates do not include a risk estimate and only include optimism bias at 5%. This is a low allowance given that both schemes are at an early stage of design. However this approach was adopted because these are Do-Minimum (DM) costs which will be avoided in the Do- Something (DS) and can therefore be offset against the DS costs. This means that these low allowances act to produce a more conservative assessment of the DS options' performance. To increase the optimism bias allowance on these DM estimates would increase the scale of DM costs which are offset against the DS costs and therefore reduce the net cost included in the economic assessment (and so counteract some of the optimism bias added to the DS costs). back [22]
23. COBA Manual: DMRB, Volume 13, Section 1, Part 2, The Valuation of Costs and Benefits:Chapter 9, Highway Maintenance Highways Agency May 2002 back [23]