Wednesday, 6 November 2013

TRAFFIC CHAOS PREDICTED IN EARLEY: A DETAILED ANALYSIS & REPORT


This paper has been produced by Dave Green at the request of the Focus editor. Dave has been a Lower Earley resident for 28 years and, until his recent retirement, he was a well qualified and experienced Civil Engineer who has specialised in Local Authority Highways and Transport fields for the best part of 40 years. He has worked directly both as a local government officer and, for the last 18 years of his career, for major consultancy companies at levels up to and including Director of Transportation. He has supported many local highway authorities especially within South East England and, in particular, he knows Berkshire’s transport networks well. He has serious concerns about the way in which future transport needs across Central Berkshire are currently being predicted and addressed. The views expressed are his own and are based on his extensive professional experience and success in these particular fields. All comments made are intended to be objective, informative and without bias. The paper is based largely upon information gleaned from the WBC web site and public meetings but Dave also relies upon some material and data produced by various community groups in Shinfield and Arborfield.

[Please note: a downloadable version of this paper is available from http://go.earleylibdems.org/traffic]

Most local readers will be well aware of just how busy our local roads are and many of you will have personal experience of trying to get in or out of Earley during typically congested morning and evening peak periods....whether on foot or bike, on a bus or in a car. However not everyone will be aware that things are likely to get much worse over the next few years because of planned and ongoing major development as well as natural traffic growth. The following is a list of just some of those developments which may have most impact on Earley roads:-


  • IQ Winnersh...ongoing redevelopment of the Winnersh Triangle site to provide around 10,000 jobs when complete.
  • Shinfield ... approved plans for 3,000 new homes and approximately 4,000 new jobs at the Science Park.
  • Arborfield... advanced proposals for 3,500 new homes and some jobs on the Garrison site.
  • Green Park and Reading International Business Parks...an estimated 14,000 jobs when full.
  • Tesco Distribution Centre on the old Courage Brewery site.
  • Ongoing major redevelopment proposals at Worton Grange and at Station Hill in Reading. 


There are also ongoing plans for major housing developments throughout the region but particularly in North and South Wokingham. The above list is not intended to be exhaustive but serves to indicate the scale of the potential problem facing our local roads.

Many of these developments are to be welcomed overall as they will bring vastly enhanced employment opportunities for our families. However, it is clear that Earley lies pretty much at the heart of all this future growth pressure which will inevitably generate far more traffic. As our local roads provide key links to Wokingham, Central Reading, Woodley and Winnersh as well as to the national road network at M4 junctions 10 and 11 it is clear for all to see that Earley’s roads will witness considerably increased traffic demand.

The Government appears to agree that we will see major traffic growth in the next few years as the Department for Transport recently published its new forecasts predicting an average increase in traffic across UK roads of 19% by 2025 and 43% by 2045. The report also implies that areas enjoying higher than average levels of economic activity and development ( such as Central Berkshire) may well see correspondingly higher than average traffic growth!

So what is the Wokingham Borough Council approach to further transport provision to meet these future demands? So far the Council has adopted a stance of providing only local improvements to facilitate access to each development from the local highway network. For example the IQ Winnersh development funded signalisation of the previous roundabouts adjacent to the site but relatively little else particularly to the key congestion hotspot at Loddon Bridge Roundabout. Similarly, the Shinfield development will fund a new and largely single carriageway Shinfield Relief Road with a new dual carriageway M4 crossing but terminating in a relatively unchanged Black Boy Gyratory. This offers no relief to existing congestion either on the Earley Outer Peripheral Route (both towards M4J11 and Lower Earley) or on the Shinfield Road towards Reading through Shinfield Rise and these existing limitations will provide severe practical limitations on future junction performance.

The Council’s approach to Transport Planning is supported by its Wokingham Strategic Transport Model (WSTM). However, because of the counter-intuitive nature of the approach, given the presently congested local highway network, available documentation on the model has been examined more closely and is discussed briefly below.


Base Year Model

The published Local Model Validation Report (LMVR) reports validation of the model version 3 (WSTM3) in accordance with government guidelines for models of its type and, by implication, intended use. However, the validation remains relatively coarse, and it is important to look at what is validated and what is not.

Individual junction turning movements are not subject to the same level of validation as links, and this is attributed in the LMVR to the turning counts being a snapshot of activity rather than a picture built up over a period. This does not, it is said in the Validation Report, invalidate the model but, in reality, neither does it help to build any confidence in it.

Journey time surveys are reported in the LMVR, and these are used for comparison with the modelled equivalents. What is clear from the comparison graphs in Appendix H of the LMVR, especially for the peak hours, is that the modelled profiles are generally much smoother than the observed. This suggests that the model fails to attribute sufficient delay to key junctions, and that this delay is swallowed up in an adjusted link transit time. In any event, there appears to be too little distinction between travel times on links and time spent queuing. The result of this is that, whilst the overall measure of journey time may be acceptable (within +/-15% of observed), the split between link transit time and queuing delay may be incorrect. The effect of this is that the model may not be relied upon to report delays at individual junctions. In addition +/- 15% of observed is in itself a marked potential deviation from reality, especially as the primary basis for future decision making, without some allowance being made for the potential for the actual numbers being significantly different.

Models are often perceived to overestimate traffic capacity simply because they do not reflect a given individual’s personal experience of delays, often gained at the same time each weekday, and generally recalled in terms of the worse end of the congestion spectrum experienced. However, the WSTM3 base year model appears to overestimate capacity significantly at some key junctions. For example the Arborfield Cross junction is reported as having at least 50% spare capacity on all approaches in both peak hours when in reality the junction already experiences marked peak period delay and congestion.

Other modelled junctions appear to be similarly affected, including the Lower Earley Way/Mill Lane junction. As many readers will know the morning peak queue at this site when travelling towards Loddon Bridge is generally measured in hundreds of metres , and the associated delay is several minutes. Modelled delays for 2010 are not published in the LMVR, but the forecasting work apparently shows that this junction does not appear to be overloaded even in 2026 with all development traffic added! One of the problems is that delays are reported as the average per vehicle for the junction as a whole, which tends to dilute the larger numbers on seriously delayed approaches and so masks the most serious problem junctions.
A shortage of time and access to data has precluded a detailed examination of every modelled junction but it is assumed that the examples quoted above and their problems may be typical of the methodology as a whole.

Absence of update: At a meeting earlier in the year, officers explained that a review was in progress of future transport problems that would not be alleviated by the provision of improvements associated with the Strategic Development Locations (SDLs). A key part of this process appeared to be a planned update of the base year model in order to move it on from 2010. The updated model has not been published, so it is assumed that it has not been developed.

Meanwhile, the data upon which the 2010 model is based is ageing steadily and must soon be regarded as too old for the purpose.

Forecasting Methodology: The basis for calculating development traffic is to take the number of land-use units (dwellings, school places, 100sqm of employment floor space, etc) and multiply by the relevant trip rate. Many years ago, weary of constant disagreements with developers, local authorities in the South East got together to create a database of trip rate surveys across the spectrum of land uses. The resulting system is called TRICS (originally Trip Rate Information Computer System), and it has become the national standard source for trip generation data. The database is continuously updated with new surveys carried out by member organisations. Data can be selected from the database on a variety of criteria – location, development size, transport characteristics and so on and the user has complete freedom to select a rate from the resulting, filtered data offered. WBC used selected TRICS data when choosing residential trip rates to attach to the SDLs in their traffic model which are considered unrealistically low given the specific circumstances of Wokingham Borough (a high car dependency , a relatively poorly developed bus network and a generally affluent population and the out of town locations of many proposed developments.

WBC use a morning peak hour trip rate for dwellings of 0.53 car trips per house and 0.28 car trips per flat. In order to test this assumption traffic was surveyed independently at Penrose Park near Arborfield Garrison between 0800 and 0900 on Tuesday 4 June 2013 (not a seasonal peak for residential trip generation), where there are 69 flats and 231 houses. Using the WBC trip rates, this much development would generate 141 trips. The surveyed total was 186, some 32% higher. Penrose Park has yet to mature as a development – as the families there grow, traffic generation will tend to increase, as it has elsewhere. Mill Lane, Lower Earley, for example is reported in TRICS at a trip rate of 0.885 (closer to the more commonly used figure of 0.85). Goldsworth Park, a mature development of almost 3000 homes in Woking, is reported at 1.016. Use of these rates would give 67% and 92% more traffic generation than the rate adopted by WBC.
Further trip rate surveys were carried out on Thursday 13 June 2013 by members of the Arborfield and Barkham Neighbourhood Plan Transport Working Group, including a repeat of the earlier survey at Penrose Park. The results show a weighted average trip rate for houses, as distinct from flats, of 0.766 trips per house, some 45% over the rates used in WSTM3. Crucially, the surveyed rate for houses at Sheerlands Road, where the development has had a little more time to mature, was 0.95.

Housing mix will, of course, affect overall trip rates but 0.53 as a garden gate rate in the specific circumstances of Wokingham Borough is clearly much too low. The model should incorporate a residential trip generation rate of not less than 0.8 for houses and the corresponding rate for flats. This would quite clearly have a considerable impact on the forecast of development traffic at all housing developments and on Earley’s roads in particular. The consequences of using the wrong trip rates for any new development are potentially severe:-

  • The additional traffic impact of new housing will be underestimated by at least 50%
  • The area over which the impact is felt will be greater than estimated
  • The developers will not pay for all the improvements needed to cope with their traffic
  • Junctions that the Council think will operate normally will actually be overloaded
  • Some of the traffic relief that the Council thinks will be brought by the proposed new infrastructure will not materialise
  • The Council (as Highway Authority) will be liable for the expense of addressing these errors at some future date

Trip rates for employment and all other land uses have not been examined in detail at this stage because of a shortage of time and so cannot be assumed to be reliable. However, the housing rates alone are sufficient to give rise to major concern about the adequacy of model outputs.

Smarter Travel Choices: WBC rely on studies carried out in Darlington, Peterborough and Worcester, between 2005 and 2008, as the basis for reducing already inadequate forecast future trip rates by a further 7.8%, founded on research carried out during the course of the Smarter Choices project. This involved considerable government expenditure on a wide variety of travel planning (personal, workplace, school, etc), a good deal of marketing, information and training, and travel awareness campaigns. The characteristics of the project sites are quite different from those at the SDL locations, especially at Shinfield and Arborfield, and a good deal of what was done in the project towns is not proposed here.
  • Darlington: confined to the urban area, with a population just under 100,000. Darlington was selected because its residents made a high proportion of short journeys and because travel patterns appeared conducive to sustainable travel interventions. Darlington had (and has) lower than average car ownership (with 69% of households owning a car) and relatively high levels of bus patronage. About three-quarters of the Darlington population had jobs based in Darlington. Wage levels were low and unemployment above the national average. In 2005, Darlington also received £1.5m as one of the Cycling Demonstration Towns, resulting in the development of seven new radial cycling routes into the town centre.
  • Peterborough: designated as a New Town in 1968, the travel choices programme focused on the urban area, with a population of 140,540. New Town status had resulted in the development of a large network of off-road cycleways. Car ownership was average (74% of households owning a car). Peterborough is located on the East Coast Main Line railway and there was a large flow of commuters travelling to London by train from the centrally located railway station.
  • Worcester: the programme of interventions was limited largely to the City of Worcester with some promotional activity expanded into other areas. The City has two railway stations and some park and ride facilities. The population of 93,000 has slightly above average car ownership (77% of households own a car).
  • Total investment in all three towns over the duration of the experiment was just over £15.5m. All three urban areas have well-developed networks of bus services. There are no published follow-up studies to determine whether the observed reductions in trip-making have been sustained since funding for continuing programmes of interventions has ended.

Compare this with the characteristics of the Arborfield Garrison site for example: 3,500 new homes; 90% of Wokingham Borough households have a car; bus services to only 2 local destinations; 46 minute journey time to Reading station plus 25 minutes to London Paddington by train; 24 minutes to Wokingham station plus 1hour 12 minutes (0756 departure) to London Waterloo by train; 95% of journeys to work going off-site; minimal proposed investment in smarter choices interventions; etc, etc. The dissimilarities in travel characteristics, and the propensity for any sort of smarter travel choices programme to have significant impacts, could hardly be greater.

It is impossible to quantify reliably how smarter choices may affect traffic arising from any of the SDLs, but the dissimilarities between the study locations and the Shinfield and Arborfield sites together with the lack of meaningful proposals to influence the choice of travel mode are unlikely to lead to any significant reduction in home-based trip making by car.
Underlying Traffic Growth: As already discussed above the most recent government forecasts of traffic growth were published by the Department for Transport (DfT) earlier in 2013. The forecasts deal with the inevitable uncertainties of economic growth, oil prices and population growth by presenting an array of possible future growth profiles. The central forecast, generally used where a single future scenario is being projected, shows more rapid growth than was forecast in 2010, when the WSTM3 was developed. The most recent DfT central forecast suggests a UK average growth of 19% by 2025 and 43% by 2045 and implies that growth in Wokingham will be higher because of higher local economic growth. These figures alone suggest that WBC needs to seriously reconsider its response to general traffic growth


Inappropriate Use of the Model

Assessment of individual junctions: As the name implies, WSTM3 was built as a strategic model, largely to underwrite the broad strategy assessments needed to develop the Core Strategy (CS). It was never, or should never have been, the intention to use the model at a local level, and certainly not to apply the model to the assessment of individual junctions. The lack of validation on junction turning movements, and the unbalanced composition of the modelled journey time data, produces a model that is simply not fit for this purpose. When it is used in this way, serious consequences can result.

Problem Identification: There are two aspects of the modelling approach to junctions that make WSTM3 unsuitable as the basis for identifying problems on the network.

The incorrect allocation of journey times between link transit times and junction delays in the base year model will carry through to the forecast models for 2017 and 2026. Moreover, the situation will deteriorate because junctions in the base year will be acceptable when, in reality, they already generate more delays than modelled. As traffic is increased, instead of becoming oversaturated and causing exponential growth in delays, these junctions simply become more heavily loaded with more modest delays accrued. If, as seems likely, this is a network-wide issue, then future levels of network performance may be considerably worse than are being forecast. The use of an average delay per vehicle averaged over each junction masks the real problems. The Lower Earley Way junction referred to above is a good example of this. Problem identification should be directed by the worst performing arm(s) if resources are to be targeted to any useful effect.

Option Appraisal: It follows, then, that the model is ill equipped to provide a sound basis for considering any option to tackle problems identified. Relying on an approach in which the shortcomings apply to all alternatives being appraised, and are therefore somehow cancelled out simply does not work.

Using the ongoing WBC work to assess options for Arborfield Relief Road as a case in point. The relative attractiveness of the alternative routes relies on an accurate portrayal of the performance of the Arborfield Cross junction. The new junctions being incorporated into the model can be designed to a consistent up to date standard, so that the conventional capacity assessment techniques will apply reasonably well. Not so the existing junction which, because it offers too little resistance to traffic by way of delays (as modelled), makes the existing route through Arborfield unduly attractive relative to a new relief road. Thus, the performance metrics used to gauge the effectiveness of each of the options will be incorrect.


Unachievable Transport Goals

The purpose of a Local Transport Plan (LTP) is to maintain and improve transport in the area referred to and which is generally the administrative area of the responsible Highway Authority. The current LTP for WBC is LTP3 and is the third generation of a plan originally introduced in 2001.
The LTP goals set out on page 2 of the document are long on process and short on outcomes. Mostly the content discusses what the authority will do rather than what it will achieve but with some exceptions. Journey times on A class roads are quoted as 2 minutes 21 seconds per mile in 2008 and would increase by 22% if planned development were to go ahead without any associated infrastructure. However all planned infrastructure improvements are essentially local in nature and effect and will have little if any wider benefit . Yet WBC is committed to restoring average journey times in 2026 on A class roads to 2008 levels although no means have been given to demonstrate just how this will be achieved. Moreover within the Core Strategy WBC is committed to

“ ............ a good transport system that will reduce congestion”

What is of greatest concern is that all transport planning and development decisions being made by WBC at present are based upon a complete reliance upon the WSTM3 model. In turn this is a high level strategic model which appears to bear little resemblance to reality and is neither well suited nor fit for these purposes.