The reason for my inquiry is related to the traffic modelling (re.
> Strategic Traffic Assessment - Modelling Results in your plan). I
> noted that the input to the analysis (figures applied to all housing
> sites) was based on the housing distribution of the Cape Road
> development in Warwick. I'm concerned that the mix of housing types
> for this development does not reflect what would be found in
> out-of-town developments, specifically those at Gallows Hill. A quick
> calculation based on the figures given on page 7 of the report gives
> an average per-household trip rate of 0.39 (for both AMPeakHour &
> PMPeakHour) - based on the housing distribution of the Cape Road
> development. However, for the proposed housing distribution for
> Gallows Hill, based on your figure - and I'm assuming all housing and
> no apartments (you didn't mention any) - then the average
> per-household AMPeakHour trip rate rises to 0.69, and the PMPeakHour
> rate to 0.79. These figures are an increase of 62% and 100%
> respectively over the figures used for the modelling exercise. When
> you take into account that this discrepancy applies to the largest
> development in the area, which connects to roads that are already at a
> standstill at peak times, it questions the validity of the whole
> modelling exercise and I would say renders it meaningless. I'd like to
> hear what your view of this is.
>
> One thing I've not found in the whole traffic strategy is any sort of
> limit to the traffic on a road. The traffic modelling exercise seeks
> to minimise traffic queuing but places no limit on it. I see queue
> lengths of 50 to almost 100 in the modelling results, but what does
> this mean? There must be a point where the amount of traffic becomes
> unacceptable in terms of delay, quality of life etc, such that it
> would be irresponsible to put plans in place knowing that this limit
> would be exceeded. Can you tell me if such a limit exists, or is it
> assumed that new developments can generate new traffic in an
> unconstrained way?
>
> I appreciate that traffic is the responsibility of the County Council;
> however, you base your plans on their input so my points are relevant
> to this process.


You say that the difference between the figures used in the model and
those that should have been used according to the housing distribution
for Gallows Hill "is not as significant as you suggest". I'm
interested in why you think my figures are inaccurate, so let me
elaborate.

As I stated previously, my calculations are based on the values
published on page 7 of the report. The average peak hour trip rate per
dwelling will be (IN + OUT rate) x % population of that housing type,
summed across each housing type present. So for Private Housing the AM
figure will be (0.19 + 0.50) x 0.16, for Private Apartments this will
be (0.02 + 0.29) x 0.54 and so on. Thus the average AMPeakHour trip
rate used in the model is (0.19+0.50)x0.16 + (0.02+0.29)x0.54 +
(0.18+0.37)x0.15 + (0.08+0.12)x0.15, which gives a result of 0.39.
Similarly the PMPeakHour trip rate is (0.57+0.27)x0.16 +
(0.17+0.05)x0.54 + (0.40+0.32)x0.15 + (0.14+0.08)x0.15 is also 0.39.

Using a distribution of Private Housing 60%, Private Aparts. 0%,
Social Housing 40%, Social Aparts. 0%, the AMPeakHour trip rate is
(0.19+0.50)x0.6 + (0.02+0.29)x0 + (0.18+0.37)x0.4 + (0.08+0.12)x0
which gives 0.64, while the PMPeakHour trip rate is (0.57+0.27)x0.6 +
(0.17+0.05)x0 + (0.40+0.32)x0.4 + (0.14+0.08)x0 which gives 0.79.

0.64 compared with 0.39 is an increase of 64%
0.79 compared with 0.39 is an increase of 102%

Perhaps you can explain what you find wrong with these calculations.
I'm also very confused over the figures you say the County Council use
in their assumptions - these are different from those in the model.
Why on earth would the model commissioned by the County Council not be
using the Council's figures?

Even if we assume that the discrepancy between the figures that were
used in the modelling were closer to yours than mine, the complexity
of the model (which is modelling an already overloaded network) is
likely to be such that you just cannot predict the consequences. Only
remodelling could show this. The Gallows Hill site is the largest in
the area, so even a small change in traffic rates is likely to have a
dominant effect on traffic. I don't think this can be so easily
dismissed.

I'm also concerned that I see no estimation of accuracy in this
modelling. As someone who started life as a scientist and then spent a
lifetime in software development, I've often tried to persuade
colleagues, particularly management, to be cautious of output of
computer calculations. At University, experimental results would be
discarded by tutors if you were not able to provide accuracy
calculations. Business computing never seems to apply this discipline,
input and output accuracy is often unquantifiable but still there is a
wish to treat the results with an accuracy that is unwarranted.
Beautifully crafted computer output can be very compelling. We used to
say GIGO - Garbage In, Garbage Out, and all the sophistication of a
complex model doesn't change the fundamentals of this; the output may
be highly crafted, fantastically accurately calculated, but still
ultimately of little value.

I also don't really understand your point where you say "Despite the
trip rate difference, we are of the view that the study is still valid
as a number of other factors will balance against the difference".
Surely the factors that balance against the difference are
independent, and have nothing to do with the validity of the study.
These other factors do not make the study valid. Additionally, while
the factors made be true, you have no idea what they are balancing
against if the model is invalid. The actual weight of traffic may
completely swamp these mitigating factors, rendering them ineffective.

Lastly, as to traffic congestion, I agree this is subjective. I still
think it would be helpful - to you guys at least - to have some
threshold to be able to judge against. Something for the future
perhaps.